Record what Kiwix returns and why the gate is not one (V-668)

gofmt on cmd/mavwaked/silero.go came in with a99932b and blocked make test.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

.gitignore

@@ -70,3 +70,5 @@ coverage.out
 
 # root .env — MAVEN_AMBIENT_TOKEN and friends, same class as deploy/telegram.env
 .env
+# silero-vad, downloaded (see AGENTS.md)
+/models/vad/

AGENTS.md

@@ -95,6 +95,22 @@ model: the code puts `query: ` in front of a question and `passage: ` in front
 of a stored note, which is how e5 was trained. The quantized file is the one
 that is downloaded, deployed and measured.
 
+## Voice activity model for mavwaked
+
+`mavwaked` decides an utterance has started with silero-vad when `-vad-model`
+points at it, and with an energy threshold when it does not. The model is 2.3MB
+and is not committed:
+
+```sh
+mkdir -p models/vad
+curl -sL -o models/vad/silero_vad.onnx \
+  https://github.com/snakers4/silero-vad/raw/master/src/silero_vad/data/silero_vad.onnx
+```
+
+It needs the same `libonnxruntime.so` the embedder needs, passed as `-onnx-lib`
+or read from `MAVEN_ONNX_LIB`. The measurement is
+`docs/evals/2026-08-09-silero-vad.md`, and the tests skip without the file.
+
 **Also need ONNX Runtime** (`libonnxruntime.so`):
 
 ```sh

CLAUDE.md

@@ -47,27 +47,63 @@ free — `worldGap` in `cmd/mavend/worldmodel.go`, which the owner hears instead
 answer. A box with no `workstation` block behaves exactly as it did before the seam: naming
 a gap requires a gap. The offload table in `docs/offload.md` says which caller is which.
 
+**Speech-to-text moved on 2026-08-09** (V-486). `sttSeam` in `cmd/mavend/voicewire.go`
+builds an `stt.Pair` beside `modelSeam`, preferring CrisperWhisper 2.0 turbo on workpc
+with mavsttd as the floor. It takes only the silent half of the rule. A worse
+transcript is still a turn, so `stt.Pair` has no `TranscribeRemote`. The fallback is
+never spoken. CW2 turbo scores **10.4% WER in Russian against 27.5%** for the `ggml-small.bin`
+mavsttd loads, over 200 Golos clips
+(`docs/evals/2026-08-09-crisperwhisper2-russian-wer.md`). It runs in Intended mode, not
+Verbatim, though that corpus cannot separate the two.
+**whisper.cpp cannot load CW2 at all.** It reads its language count off the vocabulary
+size, and CW2's 51897 tokens shift seven special token ids. So it is not a second
+endpoint on mavgpud. It is its own transformers service on port 8081
+(`deploy/cw2/serve.py`), which Maven reaches directly. `stt.HTTPTranscriber`
+posts raw PCM to it with a bearer token, because audio is the most sensitive thing that
+crosses this seam. The switch is `workstation.stt` in
+`deploy/mavend.json`, and deleting the block sends every utterance to mavsttd.
+**mavgpud runs that service as a second child.** That is not an optimisation. CW2 is a
+ROCm process on the same card, so it registers on the KFD like any contender. Under its own
+systemd unit it made mavgpud evict llama-server every few seconds. That took the
+gemma-4-12b arm down for eight minutes on 2026-08-09 before anyone noticed. The card needs
+one owner. Any GPU service added beside this daemon has the same defect, so add it to
+`cmd/mavgpud` and not to systemd. CW2 is on the yield clock and not the idle one. At 1.6GB
+it denies the card to nobody, and unloading it would only send the next voice turn to the
+homesrv floor.
+Text-to-speech has not moved and piper on homesrv is still the only synthesizer.
+
 ## Build & test
 
 CGO daemons (`mavend`, `mavsttd`, `mavttsd`, `mavenclient`) need the vendored toolchain
 and libs wired through the Makefile — **do not** call `go build` on them bare, use `make`:
 
 ```sh
-make build          # all 9 binaries
+make build          # all 11 binaries
 make build-web      # single daemon (pure-Go ones: web/waked/poll/caldav build without CGO)
 make test           # go test -race across ./internal/... ./cmd/... with CGO env set
 ```
 
-Run a single test (must carry the CGO env for packages that touch STT/TTS/voice):
+Run one package or one test with `make t`. **Do not hand-write the CGO preamble.**
+Past sessions pasted it about 390 times. That is where the shell-quoting failures
+came from. This box runs zsh, so an unquoted `-run Test*` or `--include=*.go`
+dies on "no matches found" before `go` is ever reached.
 
 ```sh
-CGO_CFLAGS="-I$(pwd)/deps/include -I$(pwd)/deps/whisper.cpp/ggml/include" \
-CGO_LDFLAGS="-L$(pwd)/deps/lib -Wl,-rpath,$(pwd)/deps/lib" \
-LD_LIBRARY_PATH="$(pwd)/deps/lib" \
-deps/go/go/bin/go test -run TestName ./internal/router/
+make t PKG=./internal/router/
+make t PKG=./cmd/mavend/ RUN=TestSimulator
+make t PKG=./internal/router/eval/ RUN='TestONNX' V=1   # V=1 for -v, RACE=0 to drop -race
 ```
 
-Pure-Go packages (`router`, `memory`, `mavweb`, …) run under a plain `go test ./pkg/`.
+`t` carries `-race`, so a green `make t` cannot turn red under `make test`. It carries
+`-count=1`, so a cached PASS from before your edit is never mistaken for a result.
+
+It also sets `MAVEN_ONNX_LIB`, which the hand-written recipe did not. The four
+`TestONNX*` measurements self-skip when that variable is unset. The run still prints
+`ok`. So every targeted eval done the old way reported the hash ratchet while reading
+as a real embedder score.
+
+Pure-Go packages (`router`, `memory`, `mavweb`, …) also run under a plain `go test ./pkg/`,
+but `make t` works everywhere and is one thing to remember.
 
 ## The daemons (`cmd/`)
 
@@ -79,22 +115,53 @@ Pure-Go packages (`router`, `memory`, `mavweb`, …) run under a plain `go test
 | `mavttsd` | Text-to-speech (piper subprocess). |
 | `mavwaked` | Wake-word / VAD gate. **Not on homesrv** — see below. |
 | `mavenclient` | Voice loop client (mic → stt → core → tts). **Not on homesrv** — see below. |
-| `mavpoll` | Telegram long-poll reach. |
+| `mavpoll` | Environment poller: netdata alarms, uptime-kuma, zenmoney, wireguard presence. Writes facts, sends nothing. Telegram is `internal/delivery/telegramsink`, not this. |
 | `mavcaldav` | CalDAV calendar sync. |
 | `mavmaild` | Mail reader (IMAP, read-only). Holds the IMAP password; core never sees it. |
+| `mavgpud` | GPU supervisor. **Runs on workpc, not homesrv** — own unit, `deploy/mavgpud.service`. Keeps llama-server loaded while the card is free (V-488). Maven never asks it for anything, it reads `/health` through `llm.Pair`. |
+| `mavupdate` | Not a daemon. Operator CLI a human runs on the box to deploy a new build. |
+
+Two more binaries have no Makefile target and are built with `go run` or `go build` when
+they are needed. Neither is deployed.
+
+| Binary | Role |
+|---|---|
+| `mavseal` | Recovery tool. Encrypts a live tmpfs working copy back to the ciphertext file when mavend was killed before `defer st.Close()` sealed it. |
+| `labelgen` | Runs the stage 0 grammars over utterances and prints JSONL, the training data for the routing heads (V-546). |
 
 Daemons are wired socket-to-socket, not linked. `internal/ipc` is the client/server wire
 protocol; the config in `deploy/mavend.json` (with `${VAR}` env expansion from gitignored
 `deploy/telegram.env`) sets socket paths, model paths, and the phraser/embedder blocks.
 
-**Seven of the nine run on homesrv. `mavwaked` and `mavenclient` do not, and that is the
-decision, not an oversight** (Vikunja #463, `docs/plans/17-where-the-voice-loop-runs.md`).
+**`docker-compose.yml` runs five: `mavend`, `mavsttd`, `mavttsd`, `mavweb`, `mavpoll`.**
+Count against compose, not against the table. Four of the nine daemons are absent, and each
+absence has a different reason.
+
+`mavmaild` and `mavcaldav` are commented out in compose, each with the reason written
+beside it: the first needs a mail account, the second a CalDAV account, and this box has
+neither. `mavcaldav` used to appear nowhere at all, which was an oversight; it became a
+recorded decision on 07-08-2026 (V-644). Two things ride on that absence and the block
+names them. Agenda questions route to `IntentQuery` at stage 0 (V-498) and the `calendar`
+query source then reads a table nobody writes. And `loop.State.CalendarBusy` is fed by the
+same facts, so the gate's "do not nag mid-meeting" is permanently false. Its password is
+read from a file (`-pass-file`, and `-render-pass-file` for the render collection), never
+taken as a flag value, which is the rule `mavpoll` and `mavmaild` follow too.
+
+**`mavwaked` and `mavenclient` are absent by decision, not oversight** (Vikunja #463,
+`docs/plans/17-where-the-voice-loop-runs.md`).
 homesrv has a microphone — it is a laptop — but it is in the wrong room, so a wake-word
 daemon there listens to nobody. They belong on a client machine where the owner is standing.
-`ipc.Dial` already takes `tcp://host:port?token=...` through the netaddr seam, so nothing
-needs building to allow it, but no such machine exists yet. **The consequence: the wake
-word and the VAD gate are covered by unit tests and by nothing else, and no amount of
-sitting at the box changes that.** Push-to-talk through `/dash` is what QA actually covers.
+
+**That machine is workpc** (owner's correction, 2026-08-05). This section used to say no
+such machine existed, which was written when the workstation was only a model host. It is
+where he sits most of the day and it has the microphone. `ipc.Dial` already takes
+`tcp://host:port?token=...` through the netaddr seam, so the two daemons need deploying,
+not building. V-515 is that deployment.
+
+Until they are deployed, **the wake word and the VAD gate are covered by unit tests and by
+nothing else**, and push-to-talk through `/dash` is what QA actually covers. Note that
+deploying them does not by itself prove a wake word: `mavwaked` gates on energy and has no
+keyword model (V-487), so the loop runs open until that lands.
 
 ## The ecosystem: Nexus, Praxis, Hexis
 
@@ -155,6 +222,19 @@ on in deploy** — this section used to say it was wired `nil`, which stopped be
 Cascade order: `stage0.go` exact-match fast-path → LLM router (when non-nil) → classifier
 fallback. Any LLM error falls through to the classifier so a turn never breaks on the model.
 
+**A stage-0 decision is slot-extracted too, since 06-08-2026** (V-572). `fillMatchedSlots`
+in `router.go` runs the stage-2 extractor over whatever a grammar built and fills only the
+slots it left empty — a matched value always wins, because the rule read a literal pattern
+and the extractor guesses. It did not run before, so `ReminderGrammar` handed the daemon
+`HasTime: false` for "напомни в 11:00 позвонить маме" and `missingFor` read the silence as
+absence and asked "Когда?". It applies to every grammar and is inert for all but the
+reminder: `Extract` fills Time, Fn and Key and nothing else, and the query, clock, agenda,
+feed, list, task and narrative rules all emit intents with no such slot. Benchmarked at
+20000x, a stage-0 query costs 3.7µs against 3.9µs before. **`Slots.Text` is deliberately not
+filled** — a grammar that left it empty meant it, and `agendaQueryBuild` hands the query
+chain the utterance itself. Fixture unchanged at 64/91, with "slots deferred to daemon"
+6 → 0.
+
 Measured on the 77-case RU fixture. **Re-measured 2026-08-02: the classifier scores 68.8%
 full accuracy at p50 16.6µs**, not the 36.8% at p50 31ms that stood here from
 `docs/evals/2026-07-31-model-bakeoff.md`. That older figure predates the stage 0 rules and the
@@ -175,11 +255,21 @@ re-run it, start a **second** llama-server on a fixed host port — the resident
 `--port 0` inside the container and no host process can reach it.
 
 **The numbers above are the homesrv floor, not the ceiling.** With the workstation up, routing
-completes through `llm.Pair` against gemma-4-12b and scores **84.4% full / 93.5% intent-only at
-p50 329ms** — better than the resident model and about 2.5× faster (`docs/evals/2026-08-02-workstation-gemma4-12b.md`,
-Vikunja #485). The workstation is never assumed up, so both sets of numbers are live. Judge a
+completes through `llm.Pair` against the model mavgpud holds, which is better than the resident
+model and about 2.5× faster. gemma-4-12b scored **84.4% full / 93.5% intent-only at p50 329ms**
+(`docs/evals/2026-08-02-workstation-gemma4-12b.md`, Vikunja #485). The workstation is never
+assumed up, so both sets of numbers are live. Judge a
 routing change against the classifier and the resident model, since those are what always answer.
 
+**The workstation runs gemma-4-E4B since 2026-08-09** (owner's call), and it is a
+step down measured the same day (`docs/evals/2026-08-09-e4b-vs-12b-routing.md`).
+Against a same-session 12B control it scores **83.3% full / 89.6% intent-only,
+destination 19/33 against 23/33, at p50 294ms against 344ms**. So it costs four
+destination cases and buys 50ms. Read destination as the finding: it names nothing
+where the 12B names `recall` or `calendar`, which is safe but walks the whole chain.
+It also has no MTP and cannot be given any here. The only `gemma4-assistant`
+draft on disk is trained against the 12B's hidden states.
+
 **The intended third engine is not a generative model** (owner's call, 05-08-2026, V-546,
 `docs/plans/18-routing-heads-on-e5-small.md`). Routing has a bounded output space, so it is
 classification, and the 118M multilingual-e5-small is already resident. Three heads on one
@@ -191,6 +281,99 @@ was a hardcode. **Fine-tune a copy of the weights.** The resident embedder backs
 recall. Training it in place couples routing accuracy to recall@1, with nothing in the
 suite to name the trade.
 
+**Two of those heads are trained as of 08-08-2026, and they are not the three
+above** (V-661, `docs/evals/2026-08-08-routing-heads-two-head.md`). Intent and
+destination share one masked mean pool. Destination scores a mean **80.8%** over
+three seeds, best **29/33 (87.9%)**. The classifier cascade scores 12/33 and the
+cascade with gemma-4-12b scores 24/33, so a 118M encoder beats the 12B teacher it
+was distilled from. Read the best run as one seed and not a headline, because one
+case is 3 points on a fixture this small.
+Recall is 15/15 and world is 5/5. Intent is 93.6% mean over three seeds. That is
+**not** comparable to the 76.0% and 84.4% those two arms scored: a softmax has no
+clarify class, so the head's fixture is the 88 cases carrying an intent.
+
+**A fourth head asks instead of guessing, same day** (V-661,
+`docs/evals/2026-08-08-clarify-head-four-head.md`). Clarify is not a value
+of intent, so a softmax cannot emit it. It is a second question over the
+same pooled vector: can Maven act on this at all. That is why the head's
+fixture was 88 cases and not 96. Over three seeds it catches **7.0 of the 8
+`want_clarify` cases and produces 2.3 false clarifies of 88**. The cascade
+today misses 1 and produces 2, so this is parity with no rules in front of
+it. Accuracy is the wrong number here and a head that never asks scores
+91.7%. Confidence is the other half. Max softmax over the intent head reads
+**0.851 where it is right against 0.604 where it is wrong**, ranking right
+above wrong in 83.4% of pairs. `Confidence: 1.0` was a hardcode, and this
+replaces it with a signal. The two are not the same signal: one says which
+intent is unclear, the other says the utterance carries too little to act
+on. **The fourth head is not free the way the third was.** Intent,
+destination and slot F1 each move down one to four points, inside the seed
+spread. `поужинал` is a false clarify on every seed, which is the same
+defect `thinSingleToken` was narrowed for on 2026-08-01.
+
+The corpus for it is generated, because every existing row is answerable by
+construction. **The router-prompt agreement filter cannot work here**, since
+`routeGrammar` has no clarify value and a generated line always agrees with
+itself. A gemma judge replaces it. The first judge called 24 of 40
+answerable rows underspecified, because it judged against a generic
+assistant rather than against Maven's contract.
+
+**Mood is cut, not deferred.** The enum describes her own reply state, not the
+speaker's emotion, and no dataset maps onto it.
+
+**A third head landed the same day** (`docs/evals/2026-08-08-slot-head-three-head.md`).
+BIO slot tags had no Maven-domain corpus, which was true of found corpora and
+false of made ones. `label_slots.py` distils spans out of gemma-4-12b under a
+GBNF closed over Maven's own five slots. A span survives only when it is a
+literal substring of the utterance, so the agreement filter costs no second
+call. 2178 spans over 1702 rows, 37 dropped, nothing unparsed. Three heads score
+intent **92.8%**, destination **82.8%** and slot span F1 **72.4%** over three
+seeds. The slot head is free: both other numbers move less than their own seed
+spread. Epoch selection reads the intent dev slice alone. Slot F1 is still
+climbing when it stops, which costs about 4 points.
+
+The MASSIVE warm-start of step 2 is worth nothing here. Stock e5-small ties it on
+intent and leads by a third of a case on destination. Nothing argues for keeping
+that step.
+
+The floor was a corpus defect and it is fixed. The first 120 floor rows carried
+one sentence shape, so the head named a destination where the fixture says walk
+the chain. Rotating six shapes took the floor 3/7 to 6/7 and destination 75.8% to
+80.8%. What is left is calendar at 3/6 on every seed, which training cannot move:
+the possessive agenda rules claim those cases at stage 0 and name nothing, so no
+label reaches the head. That is the same trade V-660 flagged and it wants the
+owner's call.
+
+**The heads run in Go and route every turn, since 08-08-2026** (V-664,
+`docs/evals/2026-08-08-routing-heads-in-go.md`). This section used to say
+nothing of it ran. `RouterHeads` in `internal/router/heads.go` loads
+`router_heads.onnx` and reads intent, destination and clarify off one forward
+pass. It is stage 0b: after the grammars, **before** the resident model, and the
+classifier is still behind both. Through the cascade it scores intent **96.9%**
+and destination **75.8%** at p50 27.9ms. That beats the gemma-4-12b cascade,
+84.4% and 72.7%, at a twelfth of its 329ms. The workstation stays the better
+phraser and is no longer the better router.
+
+Three rules around it. The **clarify head decides first**, before the intent
+threshold. It answers a different question. A thin utterance scores low
+intent by construction, so gating it cost 6 of 8 ambiguous cases. The
+**destination head is read on `IntentQuery` only**, since no other intent
+reaches `queryWalk`. And `headsThreshold` is 0.6, the measured knee: every value
+to 0.85 drops right answers and keeps the same two wrong ones.
+
+`voice.embedder.heads_path` is the whole switch. Empty, missing or unloadable
+means the heads are nil and the cascade is byte-for-byte what shipped before
+them. **It must never be pointed at `model_path`.** The resident e5-small must
+not be replaced by the fine-tuned copy. Recall depends on that file scoring
+what it scored.
+
+**The hand-written tokenizer read every long word backwards** until this task
+(`encodeWord`, `onnxembedder.go`). It cost recall@1 7.4 points and recall@3 11.1.
+Nothing caught it because seeds and queries were mangled the same way, so cosine
+survived. The heads found it. They are trained through transformers and read
+through this. The embedder id now carries a tokenizer revision
+(`@384/tok2`), so fixing the tokenizer triggers `ReembedAll` the way swapping the
+model file does. Bump `tokenizerRev` on any change to what it emits.
+
 `Confidence: 1.0` used to be hardcoded in `llmrouter.go`, so the LLM
 path could never ask for clarification (6/6 refusal cases missed on the fixture) — Vikunja
 #359. Fixed 31-07-2026 with structural signal (single-token utterance, keyless fact, act with
@@ -254,11 +437,183 @@ of the house. A demonstrative ("отметь это как сделанное")
 `h.surfacedItems` only when exactly one item was spoken. Otherwise the turn goes back to
 the cascade rather than transitioning the wrong item.
 
+**Who claimed a turn is now recorded, and so is who did not** (V-564, umbrella
+V-558). Arbitration between the claimants on the utterance stream is order,
+hardcoded in the pre-route resolver ladder, in `buildRouter` and in
+`querySources`. `internal/decision` records one `Record` per turn: every
+claimant, what it would have made the turn, the score it reported, and whether
+it won, declined, lost on score, was thinned by a gate or was **never asked**.
+The record rides the context, the same seam `querysource.go` uses, so a claim
+site cannot change a route and a context with no record costs nothing. It is
+installed in `runTurn`, so the mic, telegram and the web all leave the same
+trail. Storage is a 25-turn in-memory ring on the handler (`decision.Ring`),
+read over `ipc.TurnDecisions` and rendered as the second table on `/trace`.
+**It also persists, since 06-08-2026, and that reverses what this section used to
+say** (V-629, `docs/plans/21-persisting-the-routing-trace.md`). The old rule was
+that nothing persists, because a turn record is read minutes later or never. The
+owner reversed it: the routing heads (V-546) cannot be fitted or calibrated
+without real utterances, and 9 of the 31 modes in `internal/modes` have no seed
+example at all. The ring did not move. It is still what `/trace` reads and still
+what a test with no store gets. `cmd/mavend/routingtrace.go` is a second sink
+beside it, writing `routing_traces` (migration #23). The utterance is stored in
+clear, because a 384-dimension vector of a short sentence is substantially
+recoverable and storing vectors instead would be a privacy claim we cannot
+support. What makes it safe is the same thing that makes the fact store safe.
+Retention is 14 days, enforced on write and again on start, so a box that goes
+quiet does not keep every row. Nothing reads it outward, and the rule
+that his notes and facts are never search input covers this table. `Store.Wipe`
+deletes it with everything else. A correction (V-630) is promoted out into a
+seed-shaped row in `routing_labels` (migration #24) and kept, because a label is
+not a transcript. The transcript still expires. The gesture that writes one is
+two buttons beside the reply on `/chat`, reached over `ipc.CorrectTurn` and the
+trace id that now rides back on `ipc.ChatReply`. A turn marked wrong with no
+target is a usable negative, so naming the intent is never required. The target
+is one of the seven intents and never free text. **All three reaches offer it as
+of 06-08-2026**, and this section used to say only `/chat` did. Voice is the
+`repair` rung, which has read spoken corrections since V-455 and now writes the
+durable label beside the classifier seed it always wrote; a spoken negative with
+no target is its own rung, `repair-negative` (V-636, `docs/plans/22-correcting-a-turn.md`).
+Telegram is an inline keyboard under the reply, and it needed the chat to become
+readable first — **telegram is no longer outbound only** (V-637,
+`docs/plans/23-inbound-telegram.md`). The poller is dark unless the `telegram`
+block says `intake`, it long-polls because the box takes no inbound connections,
+it accepts `chat_id` and no other sender, and it drops whatever queued while the
+daemon was down. It reaches the daemon through `ipc.CoreAPI` alone, so a chat
+turn takes the path `POST /api/chat` takes. Note that the turn source is still
+`tap:text` for both, so provenance cannot tell a chat turn from a typed one.
+Adding a rung to the ladder
+in `runTurn` means adding its name to `preRouteLadder` in
+`cmd/mavend/decisiontrace.go`, or that rung is silently missing from the record.
+
+**A route now says where the answer lives, not only that the turn is a question**
+(V-655, 07-08-2026). `query` was a shrug. The cascade sorted an utterance into one of
+seven intents, with stage 0, the resident model and the classifier behind it. Then
+`IntentQuery` handed the turn to `querySources` in the daemon. That is twenty-two branches
+deciding by seed similarity in a fixed order. It has no fixture and no accuracy
+number, no model arm and no floor. `Decision.Source` (`internal/router/source.go`) is
+the second half of the route. Twelve destinations, not twenty-two. The three recall
+passes plus `fact-by-key` are one destination from outside. So are search, Kiwix and
+the URL reader.
+
+**`SourceUnknown` is a real value and it is the floor.** Nothing named a destination,
+so the daemon walks the whole chain. That is byte-for-byte what shipped before the
+field existed. The classifier arm names nothing, so a box whose model is down routes
+queries exactly as it did.
+
+`queryWalk` in `cmd/mavend/actions_query.go` takes sources **out** and moves none.
+That is the safety argument and it is not negotiable. The table's order is
+load-bearing. Every comment on it argues a reason between two sources, and above all
+it carries "the owner's data first, then the world". Naming `SourceWorld` does not
+send the turn outside on its own. His notes and his facts still run first, because
+they look rather than guess.
+
+**The personal boundary is the one exception and it is deliberate.** It guesses,
+so naming `SourceWorld` drops it. That is what stops it answering "кто такой
+Линус Торвальдс?" with "не нашла у тебя такой записи", which it did on
+2026-08-07. `TestNamingRecallKeepsTheBoundary` pins the other half: naming
+`SourceRecall` keeps the boundary in front of the world.
+
+**Only a stage 0 grammar may drop it** (owner's call, 09-08-2026, V-666). The
+question of who is allowed to was open until then. Three deciders name a
+destination and two of them infer it: the routing heads and the resident model.
+An inferred `SourceWorld` on a question about him would reach SearXNG, and that
+widens what is asked rather than costing a local answer. So `Decision.SourceAnchored`
+carries the provenance. It is a field and not `Stage == 0`. Stage 0 also means
+confidence 1.0 and an anchored claim band, and one of those could stop implying
+the others. `queryWalk` reads it for the source marked `boundary: true` and for
+no other. So every other guesser still comes off the turn, whoever named the
+destination. `TestOnlyAGrammarMayDropTheBoundary` pins both directions.
+`definitionQueryPattern` claims "кто такой X", so the 2026-08-07 case is still
+anchored and still answered.
+
+What comes out is only the sources that **guess**. Those decide a turn is theirs by
+cosine against frozen seeds, then answer whatever they claimed. They hold no table
+that could come back empty. Weather is the pure case and has no local data at
+all. It was measured on the box on 2026-08-07
+(`docs/evals/2026-08-07-week-of-usage.md` section 4). It answered both "что такое
+TCP?" and "сколько будет 17 на 23?" with "для какого города?". The feed answered "какой у меня любимый язык?" with kernel headlines.
+The personal boundary answered "кто такой Линус Торвальдс?" with "не нашла у тебя
+такой записи". A source that guesses is marked `guesses: true` in the table. One that
+looks is not, and it is always asked.
+
+Stage 0 fills the destination where a rule already knows it. `WorldQueryGrammars()`
+(`internal/router/worldquery.go`) claims "что такое X" and "сколько будет 17 на 23".
+It is wired after the agenda rules and **before** the feed and list rules.
+"что такое лента" is a definition question, and the feed rule would take it on the
+noun alone.
+`calendar-query` and `event-time-query` name the calendar. The possessive agenda rules
+deliberately do not. "что у меня в списке покупок" matches `agenda-query`, and naming
+the calendar there would take the list source off the turn.
+
+Fixture unchanged at **69/91 classifier+ONNX**, measured both sides. That is the
+expected result, because it scores intent and no case here changes intent.
+
+**The destination has its own fixture and its own number as of 08-08-2026**
+(V-659, `docs/evals/2026-08-08-destination-fixture.md`). This section used to say
+it had neither. `want_source` on `eval.Case` is a pointer, because the destination
+has three states and a bare string has two. Absent is every intent but query,
+which never reaches `queryWalk`. Present and empty is the `SourceUnknown`
+contract: name nothing and walk the chain. Present and named is a destination the
+route must produce. Thirty-three of ninety-six cases carry one.
+
+A destination miss does **not** fail the case. It lands in `Outcome.SourceReason`
+and never in `Reasons`, so `Accuracy` and `IntentAccuracy` mean what they meant
+and `SourceAccuracy` is a second number over the labelled cases only. Intent and
+destination are two decisions, and one number hides which one moved. A route that
+lost its intent scores no destination hit, or a clarify would satisfy an empty
+label for free.
+
+Measured classifier+ONNX: intent **73/96 (76.0%)**, destination **12/33 (36.4%)**.
+The split is the finding. World is 5/5, because a stage 0 rule names it. The
+`SourceUnknown` floor is 5/7. Calendar is 2/6, because the possessive agenda
+rules deliberately do not name it. And **recall is 0/15, because nothing
+anywhere names it**. Those turns are still answered, since the chain walks
+recall early. Recall is the number the fourth head has to move.
+
+Seven cases assert the floor and five of them are homelab operations. They
+cluster because `SourceRecall`, `SourceNetwork` and `SourceAttention` overlap on
+every question about the box. The other two are `ru-query-005` and
+`ru-query-014`. No query source reads the reminder store, and a deadline could
+sit in tasks, the calendar or Praxis. `mavpoll` writes its netdata and uptime-kuma
+observations into the fact store recall reads. That is a finding about the enum,
+not a gap in the labelling. The owner confirmed all seven floor labels on
+08-08-2026, so they are a decision rather than an agent's guess.
+
+`baselineGrammars` in `eval_test.go` mirrors `buildRouter` and had drifted:
+`WorldQueryGrammars` was wired into the daemon by V-655 and not into the mirror,
+so the fixture scored a grammar set nobody runs. Fixed by V-659, worth 3 points of
+destination and nothing else. Check that function when adding a grammar.
+
+**The model arm landed the same day** (V-660,
+`docs/evals/2026-08-08-destination-model-arm.md`). `routeGrammar` carries a
+`source` rule closed over `router.Sources` plus the empty floor, so the model
+cannot emit a destination that does not exist. The prompt lists the twelve in
+Russian and says `""` is a normal answer to give often. `LLMRouter.Route` reads it
+back through `ValidSource` and on `IntentQuery` alone. Against gemma-4-12b on the
+workstation the cascade scores destination **24/33 (72.7%)** with intent unmoved
+at 84.4%, and **recall goes 0/15 to 14/15**. The resident Qwen3-1.7B is
+unmeasured, because it binds `--port 0` inside the container.
+
+**Stage 0 now costs four destination points.** It did not before. The four cases
+the cascade loses and the model alone wins are all calendar. The possessive
+agenda rules claim them first and name nothing on purpose. That caution was free
+while nothing downstream could name anything either. It is not free now, and the
+fix is the owner's call rather than a quiet edit.
+
+The last arm is V-546. Intent, mood and BIO slot tags were already three heads on
+one forward pass of the resident e5-small. Destination is a fourth head on the
+same pass, and 72.7% from a 12B teacher is the label source for training it.
+
 ## LLM output contract
 
-All phrasing paths emit `{"response":"...","mood":"..."}` (parsed in `replier_llm.go` and
-`internal/phraser/llmphraser.go`), with fallback to plain text and the legacy
-`{"body","summary"}`. Mood is a fixed enum. Router prompt is a separate contract:
+All phrasing paths emit `{"response":"...","mood":"..."}`, with fallback to plain text when
+the model skips the JSON. **One parser, `parseResponseMood` in
+`internal/phraser/parse.go`**, and every path reaches it: the six `LLMPhraser` methods,
+`PhraseWorld`, and `Replier.PhraseReply`, which `cmd/mavend/replier_llm.go` wraps — that file
+holds the stub fallback and no parsing of its own. The legacy `{"body","summary"}` fallback
+was deleted on 2026-08-06 (V-397): it was the contract before `{"response","mood"}` replaced
+it, no prompt asks for that shape, the GBNF cannot emit it, and no test covered it.
+Mood is a fixed enum. Router prompt is a separate contract:
 `[{"intent":<enum>, key?, value?, text?, verb?}, ...]`, 7 intents (`fact, reminder,
 note, query, act, chat, system`). `llm/check_prompt_parity.py` in the training
 workspace enforces that the Go and relabelling prompts remain identical.
@@ -320,11 +675,25 @@ world questions, so she needs to read external sources. What replaces it:
   `wikipedia_ru_all_maxi_2026-02` verbatim** through `kiwix.book_ru`. The RU→EN rewriter
   is the workaround for an English book and is skipped there. Kiwix catalog names come
   from the filename, not the `<name>` field.
+  **That verbatim path sent the whole sentence to a keyword engine until 09-08-2026**
+  (V-668, `docs/evals/2026-08-09-kiwix-topic-retrieval.md`). Kiwix ranks by keyword
+  overlap, so the question words outrank the one word naming the article. "что такое TCP"
+  returned "Перехват TCP-соединения". "кто написал Войну и мир" returned an episode of
+  Doctor Who. `kiwix.Topic` drops the narrative request, the interrogative and a verb
+  behind one. `kiwix.TitlePath` tries the exact article first, since a ZIM is addressable
+  by title and a wrong title is a 404. Four of eight questions reach the right article
+  where they did not, two were already right, and nothing regressed. Both apply on the
+  verbatim path alone. The rewriter already reduces a question, and reducing twice takes
+  the topic off its input.
   `Response.Empty()` is the whole gate and there is no quality threshold in front of it:
   the three signals one could read were measured on 2026-08-05 and none of them separate a
   real question from an invented one. Token overlap would cost "столица Франции" its
   answer, because the answer is Париж and that word is not in the question. See
-  `docs/evals/2026-08-05-search-quality-signals.md` (V-539). **Which query source claimed
+  `docs/evals/2026-08-05-search-quality-signals.md` (V-539). **The embedder is not a
+  fourth signal**, measured 2026-08-09 (V-668). Query-to-passage cosine scores 0.79 to
+  0.91 on answerable questions and 0.75 to 0.84 on unanswerable ones, and the sets
+  overlap. The wrong TCP article scored 0.8653, above five of six unanswerable rows. It
+  measures topic and not whether the passage answers, so no threshold splits them. **Which query source claimed
   a turn is readable on `/chat`** as a badge beside the reply, carried on
   `ipc.ChatReply.Source` and noted by `noteQuerySource` in `cmd/mavend/querysource.go`. It
   rides the context, so `handleText` keeps the one string signature the mic, telegram and
@@ -363,8 +732,11 @@ start of a session rather than one lookup per first use:
 ToolSearch("select:mcp__vikunja__list_tasks,mcp__vikunja__get_task_details,mcp__vikunja__create_task,mcp__vikunja__update_task")
 ```
 
-`update_task` carrying a `description` resets `done` to false, so closing a task with a
-write-up takes two calls: the description, then `done: true`.
+**Close a finished task with `done: true` and nothing else** (owner's call, 07-08-2026).
+Do not write a completion summary into the description on the way out. It is lost anyway,
+and the durable record is the commit messages and the merged PR. Note that `update_task`
+carrying a `description` resets `done` to false, which is why a write-up ever took two
+calls.
 
 ## Session workflow
 

Makefile

@@ -16,7 +16,7 @@ PIPER_BIN     := $(shell pwd)/deps/piper/piper
 PIPER_MODEL   := $(shell pwd)/models/tts/ru_RU-irina-medium.onnx
 PIPER_ESPEAK  := $(shell pwd)/deps/piper/espeak-ng-data
 
-.PHONY: simulate stt-fixtures test-stt-golden all build build-stt build-tts build-daemon build-client build-waked build-web build-poll build-caldav clean test fmt-check vet run-stt run-tts run-web download-embedder deps-go deps-sentinel tidy eval-router eval-reach eval-recall eval-phrasing eval-models build-gpud
+.PHONY: t audit simulate stt-fixtures test-stt-golden all build build-stt build-tts build-daemon build-client build-waked build-web build-poll build-caldav clean test fmt-check vet run-stt run-tts run-web download-embedder deps-go deps-sentinel tidy eval-router eval-reach eval-recall eval-phrasing eval-models build-gpud
 
 all: build
 
@@ -128,6 +128,35 @@ test: fmt-check vet
 	CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
 		$(GO) test -race -coverprofile=coverage.out ./internal/... ./cmd/...
 
+# t — run ONE package or ONE test with the toolchain env already wired. This is
+# the iteration target; `test` is the gate. Reach for it instead of pasting the
+# CGO_CFLAGS/CGO_LDFLAGS/LD_LIBRARY_PATH preamble by hand, which is how it was
+# done ~390 times across past sessions and is where the shell-quoting failures
+# came from -- the interactive shell here is zsh, and an unquoted `-run Test*`
+# or `--include=*.go` dies on "no matches found" before go ever starts.
+#
+#   make t                                     # whole tree (same scope as `test`)
+#   make t PKG=./internal/router/
+#   make t PKG=./cmd/mavend/ RUN=TestSimulator
+#   make t PKG=./internal/router/eval/ RUN='TestONNX' V=1
+#   make t PKG=./internal/store/ RACE=0        # drop -race when iterating hot
+#
+# -race is on by default so a green `make t` cannot turn red under `make test`.
+# -count=1 because a cached PASS from before your edit is worse than no answer.
+# MAVEN_ONNX_LIB is set for the same reason: the four TestONNX* measurements
+# self-skip when it is unset, so a targeted eval run would otherwise report the
+# deterministic hash ratchet and look like it scored the real embedder.
+PKG  ?= ./internal/... ./cmd/...
+RUN  ?=
+V    ?=
+RACE ?= 1
+
+t:
+	CGO_CFLAGS="$(CGO_CFLAGS)" CGO_LDFLAGS="$(CGO_LDFLAGS)" LD_LIBRARY_PATH="$(shell pwd)/deps/lib" \
+	MAVEN_ONNX_LIB="$(MAVEN_ONNX_LIB)" \
+		$(GO) test $(if $(V),-v,) $(if $(filter-out 0,$(RACE)),-race,) -count=1 \
+		$(if $(RUN),-run '$(RUN)',) $(PKG)
+
 # eval-router — score the held-out RU routing fixture (internal/router/eval).
 # Verbose so the report tables land in the terminal. MAVEN_ONNX_LIB points the
 # prod-representative baseline at the vendored runtime; override it or set it
@@ -189,6 +218,16 @@ eval-models:
 # scores the fixtures against ggml-small and self-skips when the model is
 # absent, and TestGoldenFixturesAreCanonical, which checks the committed audio
 # and the manifest with no model at all.
+# audit — the repo inventory: LOC per package, open TODOs, real stubs, living-doc
+# staleness, test shape, packages with no test. Read-only, prints, writes nothing.
+# Run it instead of rebuilding the same greps by hand; past sessions spent 93 of
+# them on this before their first edit. SECTION=loc|todo|stubs|docs|tests|gaps
+# narrows it.
+SECTION ?= all
+
+audit:
+	@SECTION="$(SECTION)" ./scripts/audit.sh
+
 stt-fixtures:
 	./scripts/gen-stt-fixtures.sh
 

cmd/labelgen/main.go

@@ -0,0 +1,113 @@
+// Command labelgen labels utterances with the stage 0 grammars and prints JSONL.
+//
+// docs/plans/18-routing-heads-on-e5-small.md calls the labeled set the whole
+// project, and it names the stage 0 grammars as the high-precision label
+// functions to start from. This runs them — the real ones, in the real
+// buildRouter order — rather than a reimplementation, so a rule change moves
+// the training data with it.
+//
+// A grammar that declines leaves the line unlabeled. Those go to the model, and
+// keeping them is the point: a set labeled only by the rules teaches only the
+// rules.
+//
+//	go run ./cmd/labelgen < utterances.txt > labeled.jsonl
+//
+// The wakeword-act grammar is absent, because its allowlist is the deployment's
+// enabled tool names and this tool has no deployment. Every other rule is here.
+package main
+
+import (
+	"bufio"
+	"encoding/json"
+	"fmt"
+	"os"
+	"strings"
+
+	"github.com/kami/maven/internal/router"
+)
+
+// label is one output row. The grammar name rides along so a reviewer can see
+// which rule made the claim, and so a rule that turns out to be wrong can have
+// its rows pulled without re-running everything.
+type label struct {
+	Utterance string `json:"utterance"`
+	Intent    string `json:"intent,omitempty"`
+	Grammar   string `json:"grammar,omitempty"`
+	Key       string `json:"key,omitempty"`
+	Value     string `json:"value,omitempty"`
+	Fn        string `json:"fn,omitempty"`
+	Text      string `json:"text,omitempty"`
+	Labeled   bool   `json:"labeled"`
+}
+
+// grammars mirrors buildRouter's order in cmd/mavend/voicewire.go. Order is
+// load-bearing there and so it is here: the agenda rules must sit after the
+// clock rules, Praxis before the capture marker, the narrative rules last.
+func grammars() []router.Grammar {
+	var g []router.Grammar
+	g = append(g, router.SystemTimeDateGrammars()...)
+	g = append(g, router.AgendaQueryGrammars()...)
+	g = append(g, router.FeedQueryGrammar())
+	g = append(g, router.TaskListGrammar())
+	g = append(g, router.ListGrammars()...)
+	g = append(g, router.ReminderGrammar())
+	g = append(g, router.PraxisGrammars()...)
+	g = append(g, router.TaskCaptureGrammar())
+	g = append(g, router.NarrativeQueryGrammars()...)
+	return g
+}
+
+func match(gs []router.Grammar, utterance string) label {
+	out := label{Utterance: utterance}
+	for _, g := range gs {
+		m := g.Pattern.FindStringSubmatch(utterance)
+		if m == nil {
+			continue
+		}
+		d, ok := g.Build(m)
+		if !ok {
+			continue // the rule saw its shape and declined it
+		}
+		out.Intent = string(d.Intent)
+		out.Grammar = g.Name
+		out.Key = d.Slots.Key
+		out.Value = d.Slots.Value
+		out.Fn = d.Slots.Fn
+		out.Text = d.Slots.Text
+		out.Labeled = true
+		return out
+	}
+	return out
+}
+
+func main() {
+	gs := grammars()
+	in := bufio.NewScanner(os.Stdin)
+	in.Buffer(make([]byte, 0, 64*1024), 1024*1024)
+	out := bufio.NewWriter(os.Stdout)
+	defer out.Flush()
+
+	enc := json.NewEncoder(out)
+	var seen, labeled int
+	for in.Scan() {
+		line := strings.TrimSpace(in.Text())
+		if line == "" || strings.HasPrefix(line, "#") {
+			continue
+		}
+		seen++
+		l := match(gs, line)
+		if l.Labeled {
+			labeled++
+		}
+		if err := enc.Encode(l); err != nil {
+			fmt.Fprintln(os.Stderr, "labelgen:", err)
+			os.Exit(1)
+		}
+	}
+	if err := in.Err(); err != nil {
+		fmt.Fprintln(os.Stderr, "labelgen:", err)
+		os.Exit(1)
+	}
+	// Coverage on stderr, so the count is visible without polluting the JSONL.
+	fmt.Fprintf(os.Stderr, "labelgen: %d/%d labeled by %d grammars\n", labeled, seen, len(gs))
+}

cmd/mavcaldav/main.go

@@ -50,10 +50,10 @@ func run(args []string) error {
 	socket := fs.String("socket", "", "core IPC socket path (required)")
 	url := fs.String("url", "", "CalDAV calendar URL, e.g. http://localhost:5232/kami/personal (required)")
 	user := fs.String("user", "", "CalDAV basic-auth username (required)")
-	pass := fs.String("pass", "", "CalDAV basic-auth password (required)")
+	passFile := fs.String("pass-file", "", "file holding the CalDAV basic-auth password (required — never passed as a flag value)")
 	renderURL := fs.String("render-url", "", "CalDAV collection maven publishes her own reminders to; empty disables rendering")
 	renderUser := fs.String("render-user", "", "basic-auth username for -render-url (defaults to -user)")
-	renderPass := fs.String("render-pass", "", "basic-auth password for -render-url (defaults to -pass)")
+	renderPassFile := fs.String("render-pass-file", "", "file holding the password for -render-url (defaults to -pass-file)")
 	renderDur := fs.Duration("render-duration", calendar.DefaultReminderDuration, "how long a rendered reminder occupies")
 	interval := fs.Duration("interval", 5*time.Minute, "poll cadence")
 	timeout := fs.Duration("timeout", 10*time.Second, "per-request HTTP timeout")
@@ -63,13 +63,22 @@ func run(args []string) error {
 	if *socket == "" {
 		return fmt.Errorf("-socket is required")
 	}
-	if *url == "" || *user == "" || *pass == "" {
-		return fmt.Errorf("-url, -user, -pass are required")
+	if *url == "" || *user == "" || *passFile == "" {
+		return fmt.Errorf("-url, -user, -pass-file are required")
 	}
 	if err := checkRenderTarget([]string{*url}, *renderURL); err != nil {
 		return err
 	}
 
+	// The password is read from a file, never taken as a flag value: an argv
+	// secret is visible in `ps` to every user on the box and lands in the compose
+	// file and the shell history. Same rule mavmaild and mavpoll follow. Read
+	// once at start, so a rotated password means a restart.
+	pass, err := readSecret(*passFile)
+	if err != nil {
+		return err
+	}
+
 	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
 	defer stop()
 
@@ -85,17 +94,20 @@ func run(args []string) error {
 		http: hc,
 		url:  strings.TrimRight(*url, "/"),
 		user: *user,
-		pass: *pass,
+		pass: pass,
 	}
 
 	var rend *renderer
 	if *renderURL != "" {
-		ru, rp := *renderUser, *renderPass
+		ru, rp := *renderUser, pass
 		if ru == "" {
 			ru = *user
 		}
-		if rp == "" {
-			rp = *pass
+		if *renderPassFile != "" {
+			rp, err = readSecret(*renderPassFile)
+			if err != nil {
+				return err
+			}
 		}
 		rend = newRenderer(core, hc, *renderURL, ru, rp, *renderDur)
 		log.Printf("mavcaldav: rendering reminders to %s", *renderURL)
@@ -131,6 +143,21 @@ func run(args []string) error {
 // It takes the whole read set, not one URL. The guarantee in the package
 // comment is about every calendar maven reads, and a second read target added
 // later must not quietly fall outside the check.
+// readSecret reads one credential from a file and refuses an empty one. An
+// empty file is a deployment mistake, not a password, and CalDAV basic auth
+// would send it and get a 401 every poll.
+func readSecret(path string) (string, error) {
+	raw, err := os.ReadFile(path)
+	if err != nil {
+		return "", fmt.Errorf("read password file: %w", err)
+	}
+	secret := strings.TrimSpace(string(raw))
+	if secret == "" {
+		return "", fmt.Errorf("password file %s is empty", path)
+	}
+	return secret, nil
+}
+
 func checkRenderTarget(readURLs []string, renderURL string) error {
 	if renderURL == "" {
 		return nil
@@ -188,6 +215,11 @@ func (p *poller) pollOnce(ctx context.Context) {
 	}
 }
 
+// maxResponseBody bounds every CalDAV response this daemon reads (the poller's
+// GET and the renderer's PROPFIND) — a misbehaving or malicious server gets a
+// truncated read, not an unbounded one.
+const maxResponseBody = 4 << 20
+
 // fetchEvents GETs the calendar URL and parses VEVENTs from the iCal response.
 func (p *poller) fetchEvents(ctx context.Context, now time.Time) ([]calendar.Event, error) {
 	req, err := http.NewRequestWithContext(ctx, http.MethodGet, p.url, nil)
@@ -203,7 +235,7 @@ func (p *poller) fetchEvents(ctx context.Context, now time.Time) ([]calendar.Eve
 	}
 	defer resp.Body.Close()
 
-	body, err := io.ReadAll(io.LimitReader(resp.Body, 4<<20))
+	body, err := io.ReadAll(io.LimitReader(resp.Body, maxResponseBody))
 	if err != nil {
 		return nil, err
 	}

cmd/mavcaldav/main_test.go

@@ -5,12 +5,38 @@ import (
 	"fmt"
 	"net/http"
 	"net/http/httptest"
+	"os"
+	"path/filepath"
 	"testing"
 	"time"
 
 	"github.com/kami/maven/internal/ipc"
 )
 
+// The password comes from a file so it never reaches argv. An empty or missing
+// file must fail at start rather than authenticate as "" against his calendar.
+func TestReadSecret(t *testing.T) {
+	dir := t.TempDir()
+	good := filepath.Join(dir, "ok")
+	if err := os.WriteFile(good, []byte("  hunter2\n"), 0o600); err != nil {
+		t.Fatal(err)
+	}
+	if got, err := readSecret(good); err != nil || got != "hunter2" {
+		t.Fatalf("readSecret(good) = %q, %v; want \"hunter2\", nil", got, err)
+	}
+
+	empty := filepath.Join(dir, "empty")
+	if err := os.WriteFile(empty, []byte("\n \n"), 0o600); err != nil {
+		t.Fatal(err)
+	}
+	if _, err := readSecret(empty); err == nil {
+		t.Fatal("readSecret(empty) = nil error, want refusal")
+	}
+	if _, err := readSecret(filepath.Join(dir, "absent")); err == nil {
+		t.Fatal("readSecret(absent) = nil error, want refusal")
+	}
+}
+
 type fakeCore struct {
 	ipc.UnimplementedCoreAPI
 	facts    map[string]ipc.Fact // composite key "key|source" → Fact

cmd/mavcaldav/render.go

@@ -142,7 +142,7 @@ func (r *renderer) listPublished(ctx context.Context) ([]int64, error) {
 		return nil, err
 	}
 	defer resp.Body.Close()
-	raw, err := io.ReadAll(io.LimitReader(resp.Body, 4<<20))
+	raw, err := io.ReadAll(io.LimitReader(resp.Body, maxResponseBody))
 	if err != nil {
 		return nil, err
 	}

cmd/mavenclient/main.go

@@ -31,15 +31,11 @@ import (
 	"errors"
 	"flag"
 	"fmt"
-	"io"
 	"log"
-	"net"
 	"os"
 	"os/signal"
 	"path/filepath"
-	"strconv"
 	"syscall"
-	"time"
 
 	"github.com/kami/maven/internal/audio"
 	"github.com/kami/maven/internal/voice"
@@ -155,9 +151,3 @@ func writeWAV(path string, a audio.Audio) error {
 // jsonUnmarshal — kept local rather than pulling encoding/json into main.go
 // top-level space.
 func jsonUnmarshal(b []byte, v any) error { return json.Unmarshal(b, v) }
-
-// keep strconv + io + net + time alive for future duration/size helpers.
-var _ = strconv.Atoi
-var _ io.Reader = (io.Reader)(nil)
-var _ = net.IPv4
-var _ = time.Second

cmd/mavend/actions_act.go

@@ -60,6 +60,17 @@ func (h *reactiveHandler) actionAct(ctx context.Context, dec router.Decision) st
 			phrase := actPhrase(dec.Slots.Fn, dec.Slots.Args)
 			h.park(dec.Slots.Fn, dec.Slots.Args, phrase)
 			return phraser.A(phraser.ActConfirm, map[string]string{"name": phrase})
+		case errors.Is(err, tool.ErrUnknownTarget):
+			// The verb reached a tool and the tail did not reach a target, so
+			// nothing ran. Saying which word she could not place is the whole
+			// answer: he either renames it or gives the row an alias that
+			// carries the target, and both are one turn away (V-634).
+			word := ""
+			var unknown *tool.UnknownTargetError
+			if errors.As(err, &unknown) {
+				word = unknown.Target
+			}
+			return phraser.A(phraser.ActUnknownTarget, map[string]string{"name": word})
 		case errors.Is(err, tool.ErrNeedsAuthedSurface):
 			// Irreversible (internal/tool/risk.go). A confirm turn would not
 			// help: everything that proposed this act — the STT, the router,

cmd/mavend/actions_note.go

@@ -9,9 +9,24 @@ import (
 	"github.com/kami/maven/internal/router"
 )
 
+// nothingToCorrectReply — what she says to a correction that points at
+// nothing. Filing it would put a sentence in his memory that reads as a fact.
+const nothingToCorrectReply = "не поняла, что поправить. скажи целиком, и я запишу."
+
 // actionNote handles router.IntentNote: embed the note, persist it, and
 // index it for recall.
+//
+// The stored body is dec.Utterance and nothing else (V-576). It is not
+// Slots.Text, not phraser output and not any other model string: a note is
+// durable, the embedder indexes it, and it comes back later as recall in his
+// own words. Phrasing belongs in the spoken confirmation.
 func (h *reactiveHandler) actionNote(ctx context.Context, dec router.Decision) string {
+	// A correction with no referent. Everything that could own one has already
+	// run by here: clarify, confirm and repair are all resolved before routing,
+	// so a fragment reaching the note path has nothing behind it (V-576).
+	if correctionFragment(dec.Utterance) {
+		return nothingToCorrectReply
+	}
 	// An utterance that explicitly files a task is work, not recall, and
 	// belongs in the task store (Vikunja #130). Checked before the embedding
 	// is paid for. Everything else is a note, exactly as before.

cmd/mavend/actions_query.go

@@ -11,7 +11,9 @@ import (
 	"unicode"
 
 	"github.com/kami/maven/internal/crawl"
+	"github.com/kami/maven/internal/decision"
 	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/kiwix"
 	"github.com/kami/maven/internal/memory"
 	"github.com/kami/maven/internal/morning"
 	"github.com/kami/maven/internal/phraser"
@@ -58,6 +60,35 @@ type querySource struct {
 	// sources search text with no notion of a day. When one of them grows a
 	// date parameter, flip its flag here.
 	dateAware bool
+
+	// dest — the destination this source serves, when the cascade named one
+	// (V-655). Several sources share a destination: the three recall passes and
+	// the fact-by-key lookup are all SourceRecall, because which of them lands
+	// the hit is an ordering detail no utterance can name. A source with no
+	// dest is reachable only by walking the chain.
+	dest router.Source
+
+	// guesses — this source decides whether the turn is its own by scoring the
+	// utterance against frozen seeds, rather than by looking something up and
+	// coming back empty.
+	//
+	// The distinction is the whole point of the field. A source that looks can
+	// be wrong about relevance and still harmless, because the miss shows up as
+	// no rows. A source that guesses answers whatever it claims: weather has no
+	// local table to miss against, so "что такое TCP?" became "для какого
+	// города?". So when the cascade names a destination, the guessers that were
+	// not named do not get to try. The lookups still run, because a named
+	// destination is evidence and not a promise.
+	guesses bool
+
+	// boundary — dropping this source widens what leaves the box, so only a
+	// literal pattern may do it (V-666, owner's call of 2026-08-09).
+	//
+	// Every other guesser costs an answer when it is wrongly taken off a turn.
+	// This one costs the rule that a question about him never reaches an
+	// upstream engine. A grammar read the words to name a destination. A model
+	// and a softmax both inferred one, and neither may spend that.
+	boundary bool
 }
 
 // querySources is the ordered chain actionQuery walks; first source to claim
@@ -66,85 +97,85 @@ type querySource struct {
 // gate was never the bug. Adding a source (Kiwix, RSS, crawler, email) is one
 // line here plus its method; where you put the line is the whole decision.
 var querySources = []querySource{
-	{name: "fact-by-key", answer: (*reactiveHandler).queryFactByKey},
+	{name: "fact-by-key", answer: (*reactiveHandler).queryFactByKey, dest: router.SourceRecall},
 	// Before "calendar" on purpose: both match "…на сегодня", and the plan is
 	// the more specific ask (its matcher requires a plan word), so the calendar
 	// listing would otherwise swallow it.
-	{name: "day-plan", answer: (*reactiveHandler).queryDayPlan},
+	{name: "day-plan", answer: (*reactiveHandler).queryDayPlan, dest: router.SourceCalendar},
 	// Also before "calendar": "что я обычно делаю по средам?" names a weekday,
 	// and the habit question is the more specific one. Its matcher requires a
 	// habit marker ("обычно", "каждый", …), so a question about this coming
 	// Wednesday still reaches the calendar.
-	{name: "habits", answer: (*reactiveHandler).queryHabits},
+	{name: "habits", answer: (*reactiveHandler).queryHabits, dest: router.SourceCalendar},
 	// Before "calendar" and before the recall sources: "что мне нужно
 	// сделать?" is a question about the task list, and the notes pass would
 	// otherwise answer it with whatever note happens to be nearest. Its
 	// matcher requires a task noun or an explicit "что … сделать", so a
 	// date-bearing question still reaches the calendar.
-	{name: "tasks", answer: (*reactiveHandler).queryTasks},
+	{name: "tasks", answer: (*reactiveHandler).queryTasks, dest: router.SourceTasks},
 	// Next to "tasks" and for the same reason: "что требует внимания?" is a
 	// question about the operational state Praxis holds, and it used to fall
 	// through every source to the web search (Vikunja #475). Its matcher needs
 	// an attention marker, and it falls through when Praxis is not configured.
-	{name: "attention", answer: (*reactiveHandler).queryAttention},
-	// Before the recall sources too: "сколько я потратил?" is a question about
-	// the money facts the poller wrote, and the notes pass would otherwise
-	// answer it from whatever he once said about spending. Its matcher needs a
-	// money noun plus an actual ask, so "я потратил весь день" is untouched.
+	{name: "attention", answer: (*reactiveHandler).queryAttention, dest: router.SourceAttention, guesses: true},
 	// Next to "tasks" and for the same reason: "что мне купить?" is a question
 	// about the shopping list, and the recall pass would otherwise answer it
 	// from an old note about the shop. Its matcher needs an explicit list
 	// marker, so "надо бы съездить в магазин" is untouched.
-	{name: "list", answer: (*reactiveHandler).queryList},
-	{name: "money", answer: (*reactiveHandler).queryMoney},
+	{name: "list", answer: (*reactiveHandler).queryList, dest: router.SourceList, guesses: true},
+	// Before the recall sources too: "сколько я потратил?" is a question about
+	// the money facts the poller wrote, and the notes pass would otherwise
+	// answer it from whatever he once said about spending. Its matcher needs a
+	// money noun plus an actual ask, so "я потратил весь день" is untouched.
+	{name: "money", answer: (*reactiveHandler).queryMoney, dest: router.SourceMoney},
 	// Also above the recall sources: "что я тебе говорил?" is a question about
 	// the facts he tapped in, and the notes pass would answer it with whatever
 	// note is nearest (Vikunja #456). Its matcher needs both halves of a
 	// history phrase and bails out when he names a topic, so "что я говорил
 	// про сервер" is still recall.
-	{name: "history", answer: (*reactiveHandler).queryHistory},
+	{name: "history", answer: (*reactiveHandler).queryHistory, dest: router.SourceRecall},
 	// Before the recall sources and before general knowledge: "что нового?" is
 	// a question about the feeds she reads, and general knowledge would answer
 	// it by inventing news. Its matcher needs a feed noun plus an ask, so
 	// "у меня новая лента в инстаграме" is untouched.
-	{name: "feeds", answer: (*reactiveHandler).queryFeeds},
+	{name: "feeds", answer: (*reactiveHandler).queryFeeds, dest: router.SourceFeeds, guesses: true},
 	// Before "calendar" and before the recall sources: "что включено дома?" is
 	// a question about the house, and the notes pass would otherwise answer it
 	// from whatever he once said about the lights. Its matcher needs a house
 	// marker plus an ask plus a device word, and it bails out on weather
 	// wording, so "какая температура на улице?" still reaches the weather
 	// source.
-	{name: "home", answer: (*reactiveHandler).queryHome},
+	{name: "home", answer: (*reactiveHandler).queryHome, dest: router.SourceHome, guesses: true},
 	// Next to "home" and for the same reason: "какие устройства в сети?" is a
 	// question about the LAN, and the recall pass would otherwise answer it
 	// from an old note about the router. Its matcher needs a network word plus
 	// an ask plus a device noun, so "интернет не работает" is untouched.
-	{name: "network", answer: (*reactiveHandler).queryNetwork},
-	{name: "calendar", answer: (*reactiveHandler).queryCalendar, dateAware: true},
-	{name: "weather", answer: (*reactiveHandler).queryWeather},
+	{name: "network", answer: (*reactiveHandler).queryNetwork, dest: router.SourceNetwork, guesses: true},
+	{name: "calendar", answer: (*reactiveHandler).queryCalendar, dateAware: true, dest: router.SourceCalendar},
+	{name: "weather", answer: (*reactiveHandler).queryWeather, dest: router.SourceWeather, guesses: true},
 	// A question about her, above the three sources that search his own data
 	// (Vikunja #555). It has no answer anywhere else: below the boundary
 	// SearXNG answers about somebody else's assistant, and above it his notes
 	// answer by proximity — "кто ты" came back from a note of his, measured on
 	// the box, because the recall index has no idea the subject is her.
-	{name: "self", answer: (*reactiveHandler).querySelf},
-	{name: "embed", answer: (*reactiveHandler).queryEmbed},
-	{name: "memory", answer: (*reactiveHandler).queryMemory},
-	{name: "notes", answer: (*reactiveHandler).queryNotes},
+	{name: "self", answer: (*reactiveHandler).querySelf, dest: router.SourceSelf, guesses: true},
+	{name: "embed", answer: (*reactiveHandler).queryEmbed, dest: router.SourceRecall},
+	{name: "memory", answer: (*reactiveHandler).queryMemory, dest: router.SourceRecall},
+	{name: "notes", answer: (*reactiveHandler).queryNotes, dest: router.SourceRecall},
 	// THE BOUNDARY. Everything above answers from his own data; everything
 	// below answers from the world's. A question about him that got this far
 	// has no answer in his data, and no outside source can supply one, so this
 	// stops the walk rather than let the encyclopedia and the model guess.
-	{name: "personal", answer: (*reactiveHandler).queryPersonal},
+	{name: "personal", answer: (*reactiveHandler).queryPersonal, dest: router.SourceRecall, guesses: true, boundary: true},
 	// The world, read live. Owner's ruling of 2026-08-02: a metasearch hit beats
 	// a frozen ZIM, so SearXNG asks before Kiwix does. Nothing of his is at
 	// stake by this point — the boundary above already stopped every question
 	// about him, and only the query string leaves the box.
-	{name: "search", answer: (*reactiveHandler).querySearch},
+	{name: "search", answer: (*reactiveHandler).querySearch, dest: router.SourceWorld},
 	// The offline encyclopedia, now the fallback for when the line is down or
 	// the search comes back empty. It reads the way it always did; what changed
 	// is that it no longer gets first refusal on a world question.
-	{name: "kiwix", answer: (*reactiveHandler).queryKiwix},
+	{name: "kiwix", answer: (*reactiveHandler).queryKiwix, dest: router.SourceWorld},
 	// LAST before the model answers from memory, and that position is the whole
 	// design (Vikunja #259): everything of his, then the search, then the ZIMs,
 	// and only then a page he named. The model does NOT come first: it
@@ -152,14 +183,70 @@ var querySources = []querySource{
 	// a 1.7B guessing at a page it cannot read is how contents get invented.
 	// This source only claims a turn where he named a URL, so it never competes
 	// with a local answer.
-	{name: "web", answer: (*reactiveHandler).queryWeb},
-	{name: "general-knowledge", answer: (*reactiveHandler).queryGeneral},
+	{name: "web", answer: (*reactiveHandler).queryWeb, dest: router.SourceWorld},
+	{name: "general-knowledge", answer: (*reactiveHandler).queryGeneral, dest: router.SourceWorld},
+}
+
+// queryWalk narrows the chain for one turn against the destination the cascade
+// named, and says which sources were left out (V-655).
+//
+// It takes sources OUT and never moves one, which is the whole safety argument.
+// The table's order is load-bearing and every comment on it argues a reason
+// between two sources; none of those reasons is about this. Above all, the
+// order carries "his data first, then the world", and a destination named by a
+// model must not be able to reverse that. Naming SourceWorld does not send the
+// turn outside — it stops the guessers from claiming it on the way.
+//
+// What comes out is exactly the sources that guess. Those decide whether a turn
+// is theirs by scoring it against frozen seeds, and then answer whatever they
+// claimed, because they have no lookup that can come back empty. That is the
+// whole of the 2026-08-07 defect: weather claiming "что такое TCP?", the feed
+// claiming "какой у меня любимый язык?", the personal boundary claiming "кто
+// такой Линус Торвальдс?". The sources that look are all still asked, so a
+// wrong destination costs nothing but the guess it prevented.
+//
+// No destination named ⇒ the table exactly as written, which is what shipped
+// before the field existed. That is the floor. The classifier arm names
+// nothing, so a box whose model is down routes queries the way it always did.
+// The personal boundary is the one exception, and anchored is what buys it
+// (V-666). A grammar matched a literal pattern to name the destination. The
+// routing heads and the resident model inferred one, and an inferred SourceWorld
+// takes the boundary off a question about him. That widens what is asked
+// upstream rather than costing a local answer, so those two keep it.
+func queryWalk(dest router.Source, anchored bool) (walk, skipped []querySource) {
+	if dest == router.SourceUnknown {
+		return querySources, nil
+	}
+	for _, s := range querySources {
+		if s.guesses && s.dest != dest && (anchored || !s.boundary) {
+			skipped = append(skipped, s)
+			continue
+		}
+		walk = append(walk, s)
+	}
+	return walk, skipped
 }
 
 func (h *reactiveHandler) actionQuery(ctx context.Context, dec router.Decision) string {
 	t := &queryTurn{dec: dec}
-	for _, src := range querySources {
+	// The roster, so the record can say which sources were never reached rather
+	// than leaving them out and letting a reader assume they looked and passed
+	// (V-564). Finish names everyone below the winner.
+	decision.Expect(ctx, decision.StageQuery, querySourceNames())
+	rec := decision.From(ctx)
+	walk, skipped := queryWalk(dec.Source, dec.SourceAnchored)
+	for _, src := range skipped {
+		rec.Note(decision.Claim{
+			Stage: decision.StageQuery, Claimant: src.name, Outcome: decision.NeverAsked,
+			Reason: "it decides by similarity and the cascade named " + string(dec.Source),
+		})
+	}
+	for _, src := range walk {
 		if dec.Continued && !src.dateAware {
+			rec.Note(decision.Claim{
+				Stage: decision.StageQuery, Claimant: src.name, Outcome: decision.NeverAsked,
+				Reason: "a continuation turn only asks the date-aware sources",
+			})
 			continue
 		}
 		if reply, ok := src.answer(h, ctx, t); ok {
@@ -172,8 +259,16 @@ func (h *reactiveHandler) actionQuery(ctx context.Context, dec router.Decision)
 			// caller asked for one, so /chat can show it (V-539).
 			log.Printf("voice: query claimed by source %q", src.name)
 			noteQuerySource(ctx, src.name)
+			rec.Note(decision.Claim{
+				Stage: decision.StageQuery, Claimant: src.name,
+				Intent: string(dec.Intent), Outcome: decision.Won,
+			})
 			return reply
 		}
+		rec.Note(decision.Claim{
+			Stage: decision.StageQuery, Claimant: src.name, Outcome: decision.Declined,
+			Reason: "it had no answer for this turn",
+		})
 	}
 	if dec.Continued {
 		// The previous question cannot be re-asked for another day. Saying so
@@ -217,7 +312,13 @@ func (h *reactiveHandler) queryFactByKey(ctx context.Context, t *queryTurn) (str
 //
 // Read-only by construction — the plan is assembled and rendered core-side and
 // nothing here schedules or announces. "что дальше?" asks for the rest of the
-// day, so that phrasing trims what has already passed.
+// day, so that phrasing trims what has already passed and reads only the next
+// morning.NextSpoken entries. Trimming alone was not enough: asked early it cuts
+// nothing, and she read 43 entries aloud in one sentence (V-618).
+//
+// "что у меня сегодня?" is a different question and is not narrowed here — it
+// carries no plan word, so IsDayPlanQuery declines it and the calendar source
+// answers the whole day.
 //
 // What surface this belongs on is still open, tracked as Vikunja #431 ("Board
 // surface: Maven holds the work board, runs the intake form, never argues").
@@ -236,16 +337,23 @@ func (h *reactiveHandler) queryDayPlan(ctx context.Context, t *queryTurn) (strin
 	}
 	// Rebuild the pure plan so the rest-of-day rendering is the same code that
 	// rendered the whole day — one formatter, one persona.
-	p := morning.Plan{Date: plan.Date}
+	//
+	// The instants are put back in the asking clock's zone on the way in. They
+	// arrive carrying whatever zone the core read them in — a calendar fact's Ts
+	// and a reminder's FireTs are UTC out of the store — and FormatRU reads the
+	// hours in the plan's own frame, so setting that frame here is what makes
+	// the recital name his clock rather than the store's (V-614).
+	zone := h.now().Location()
+	p := morning.Plan{Date: plan.Date.In(zone)}
 	for _, it := range plan.Items {
 		p.Items = append(p.Items, morning.PlanEntry{
-			At:        it.At,
+			At:        it.At.In(zone),
 			Text:      it.Text,
 			Kind:      morning.PlanKind(it.Kind),
 			Uncertain: it.Uncertain,
 		})
 	}
-	return p.After(h.now()).FormatRU(), true
+	return p.Next(h.now(), morning.NextSpoken).FormatRU(), true
 }
 
 // habitFactWindow — how many recent SELF facts the behaviour profile is counted
@@ -410,6 +518,10 @@ func (h *reactiveHandler) queryCalendar(ctx context.Context, t *queryTurn) (stri
 	return f.FormatEntries(entries, date), true
 }
 
+// homeTimeout — the whole house read. Longer than the weather call because the
+// hub is polled over the LAN and answers for every device at once.
+const homeTimeout = 10 * time.Second
+
 // queryHome answers a question about the house. Read-only by construction: it
 // calls States and nothing else, so there is no confirm turn here — the only
 // way to CHANGE something is an enabled allowlist row through tool.Executor.
@@ -426,7 +538,7 @@ func (h *reactiveHandler) queryHome(ctx context.Context, t *queryTurn) (string,
 		// unreachable case is different and homeSummary covers it.
 		return "", false
 	}
-	ctxH, cancel := context.WithTimeout(ctx, 10*time.Second)
+	ctxH, cancel := context.WithTimeout(ctx, homeTimeout)
 	defer cancel()
 	return h.home.homeSummary(ctxH)
 }
@@ -450,6 +562,10 @@ func (h *reactiveHandler) queryNetwork(ctx context.Context, t *queryTurn) (strin
 	return h.netscan.scanSummary(ctx)
 }
 
+// weatherTimeout — one geocode plus one forecast read. He asked a question with
+// a one-line answer, so a provider that is slower than this is a failure.
+const weatherTimeout = 5 * time.Second
+
 func (h *reactiveHandler) queryWeather(ctx context.Context, t *queryTurn) (string, bool) {
 	if !h.turnIsAbout(ctx, t, topicWeather, isWeatherQuery) {
 		return "", false
@@ -460,7 +576,7 @@ func (h *reactiveHandler) queryWeather(ctx context.Context, t *queryTurn) (strin
 		// so is the only honest answer; picking a city would be inventing one.
 		return phraser.Q(phraser.QueryWeatherWhere, nil), true
 	}
-	ctxWT, cancel := context.WithTimeout(ctx, 5*time.Second)
+	ctxWT, cancel := context.WithTimeout(ctx, weatherTimeout)
 	defer cancel()
 	w, err := h.weatherProvider.CurrentWeather(ctxWT, loc)
 	if errors.Is(err, weather.ErrNotConfigured) {
@@ -485,7 +601,15 @@ func (h *reactiveHandler) queryWeather(ctx context.Context, t *queryTurn) (strin
 
 // queryEmbed isn't an answer source — it's the shared cost the two recall
 // sources below both need, run once, in the position it always ran in. It
-// only claims the turn when the embedder fails.
+// never claims the turn.
+//
+// It used to claim on an embedder error, and that made a RAG hint a hard gate
+// over everything below it (V-568): one failing EmbedQuery and the memory, the
+// notes, the boundary, the search, the ZIMs, the named page and the model all
+// answered "не смогла ответить", including the questions search and Kiwix
+// would have answered without ever touching the embedder. A failed embed means
+// this source cannot claim, not that the turn is over — same shape as
+// turnVector in topics.go, which had it right.
 func (h *reactiveHandler) queryEmbed(ctx context.Context, t *queryTurn) (string, bool) {
 	// A topic source above already paid for this one; see turnVector.
 	if len(t.vec) > 0 {
@@ -493,13 +617,37 @@ func (h *reactiveHandler) queryEmbed(ctx context.Context, t *queryTurn) (string,
 	}
 	vec, err := router.EmbedQuery(ctx, h.recall.embedder, t.dec.Utterance)
 	if err != nil {
+		// Logged once, here, and the chain walks on. The two recall sources
+		// below read the empty vector and pass; the boundary drops to its
+		// offline floor.
 		log.Printf("voice: embed query: %v", err)
-		return phraser.Q(phraser.QueryFailAnswer, nil), true
+		return "", false
 	}
 	t.vec = vec
 	return "", false
 }
 
+// memoryRecallWidth and noteRecallWidth — how many candidates each recall pass
+// pulls before the gate reads them. Both are small on purpose: the gate wants a
+// best hit and its runner-up, and every further row is a margin the top match
+// has to beat.
+const (
+	memoryRecallWidth = 3
+	noteRecallWidth   = 5
+)
+
+// recallOnTopic — the topic veto both recall sources apply after the score gate
+// (#470). A memory about his slow network scored high enough to answer "почему
+// небо синее?", because the right-note and must-be-silent score ranges overlap
+// and no threshold sits between them.
+func recallOnTopic(utterance, text string) bool {
+	if memory.RecallAllowed(utterance, text) {
+		return true
+	}
+	log.Printf("voice: recall %q rejected for %q: a world question and no shared topic word", text, utterance)
+	return false
+}
+
 // queryMemory — long-term memory first: ONE search over everything Maven
 // remembers (notes and facts share this index) and ONE confidence gate, so
 // the memory that is clearly the best match answers — a note just as much as
@@ -514,7 +662,13 @@ func (h *reactiveHandler) queryMemory(ctx context.Context, t *queryTurn) (string
 	if h.recall.memStore == nil {
 		return "", false
 	}
-	hits, herr := h.recall.memStore.Search(ctx, t.vec, 3)
+	if len(t.vec) == 0 {
+		// No query vector: the embed above failed or there is no embedder.
+		// Searching on an empty vector is not a search, and its scores are not
+		// a "there is nothing" answer — pass rather than gate the chain.
+		return "", false
+	}
+	hits, herr := h.recall.memStore.Search(ctx, t.vec, memoryRecallWidth)
 	if herr != nil {
 		log.Printf("voice: memory search: %v", herr)
 		return "", false
@@ -524,12 +678,8 @@ func (h *reactiveHandler) queryMemory(ctx context.Context, t *queryTurn) (string
 		return "", false
 	}
 	text := hit.Meta["text"]
-	// The score cleared the gate and the topic still has to match (#470). A
-	// note about his slow network scored high enough to answer "почему небо
-	// синее?", because the right-note and must-be-silent score ranges overlap
-	// and no threshold sits between them.
-	if !memory.RecallAllowed(t.dec.Utterance, text) {
-		log.Printf("voice: recall %q rejected for %q: a world question and no shared topic word", text, t.dec.Utterance)
+	// The score cleared the gate and the topic still has to match.
+	if !recallOnTopic(t.dec.Utterance, text) {
 		return "", false
 	}
 	// A note is phrased in Maven's voice; a fact is read back as it was
@@ -560,10 +710,18 @@ func (h *reactiveHandler) queryMemory(ctx context.Context, t *queryTurn) (string
 // band. See memory.Confident. Failing the gate passes the turn on to general
 // knowledge, which is what "don't read back the runner-up" means here.
 func (h *reactiveHandler) queryNotes(ctx context.Context, t *queryTurn) (string, bool) {
-	notes, err := h.api.QueryNotes(ctx, t.vec, 5)
+	if len(t.vec) == 0 {
+		// Same reason as queryMemory above (V-568): with no query vector this
+		// source could not look, and could-not-look passes.
+		return "", false
+	}
+	notes, err := h.api.QueryNotes(ctx, t.vec, noteRecallWidth)
 	if err != nil {
+		// The store failed, so this source could not look either. It used to
+		// claim here, which stopped the search, the ZIMs and the model from
+		// answering a question that never needed a note (V-568).
 		log.Printf("voice: query notes: %v", err)
-		return phraser.Q(phraser.QueryFailAnswer, nil), true
+		return "", false
 	}
 	t.notes = notes
 	noteScores := make([]float64, len(notes))
@@ -573,10 +731,9 @@ func (h *reactiveHandler) queryNotes(ctx context.Context, t *queryTurn) (string,
 	if !memory.ConfidentScores(noteScores, h.recall.minScore, h.recall.minMargin) {
 		return "", false
 	}
-	// Same topic veto as queryMemory above: the best note must be about what
-	// he asked, not merely the nearest vector in the index.
-	if !memory.RecallAllowed(t.dec.Utterance, notes[0].Text) {
-		log.Printf("voice: note %q rejected for %q: a world question and no shared topic word", notes[0].Text, t.dec.Utterance)
+	// The best note must be about what he asked, not merely the nearest vector
+	// in the index.
+	if !recallOnTopic(t.dec.Utterance, notes[0].Text) {
 		return "", false
 	}
 	texts := make([]string, len(notes))
@@ -598,6 +755,23 @@ func (h *reactiveHandler) queryNotes(ctx context.Context, t *queryTurn) (string,
 // prompt, the persona block and the reply.
 const webPageContextRunes = 1500
 
+// webFetchTimeout — the whole named-page source. Longer than the other outside
+// sources because he named this page himself, so waiting for it is what he asked
+// for, and there is nothing below that can answer instead.
+const webFetchTimeout = 30 * time.Second
+
+// readBackRunes — how much of the evidence is read out when the phraser gave
+// nothing back. It is spoken aloud, so it is a couple of sentences and not a
+// page.
+const readBackRunes = 300
+
+// readBack — what an outside source says when the phraser gave nothing back.
+// The evidence is read out plainly rather than dropped, because the fetch did
+// happen and its result is a better answer than silence.
+func readBack(evidence string) string {
+	return phraser.Q(phraser.QueryFound, map[string]string{"text": crawl.TrimRunes(evidence, readBackRunes)})
+}
+
 // queryWeb — "посмотри https://example.org/x — что там?" (Vikunja #259).
 //
 // It claims a turn ONLY when he named a URL, which is what keeps a fallback from
@@ -616,7 +790,7 @@ func (h *reactiveHandler) queryWeb(ctx context.Context, t *queryTurn) (string, b
 		// guess dressed as an answer (Vikunja #479).
 		return phraser.Q(phraser.QueryPageOff, nil), true
 	}
-	ctxFetch, cancel := context.WithTimeout(ctx, 30*time.Second)
+	ctxFetch, cancel := context.WithTimeout(ctx, webFetchTimeout)
 	defer cancel()
 	page, err := h.crawler.Page(ctxFetch, link)
 	if err != nil {
@@ -637,7 +811,7 @@ func (h *reactiveHandler) queryWeb(ctx context.Context, t *queryTurn) (string, b
 	if reply == "" {
 		// No phraser (or it failed): read back the top of the page rather than
 		// pretend the fetch did not happen.
-		return phraser.Q(phraser.QueryPageText, map[string]string{"text": crawl.TrimRunes(page.Text, 300)}), true
+		return phraser.Q(phraser.QueryPageText, map[string]string{"text": crawl.TrimRunes(page.Text, readBackRunes)}), true
 	}
 	return reply, true
 }
@@ -703,7 +877,7 @@ func (h *reactiveHandler) querySearch(ctx context.Context, t *queryTurn) (string
 	if reply == "" {
 		// No phraser, or it failed. Read back the best evidence rather than
 		// pretend the search did not happen.
-		return phraser.Q(phraser.QueryFound, map[string]string{"text": crawl.TrimRunes(resp.Snippets()[0], 300)}), true
+		return readBack(resp.Snippets()[0]), true
 	}
 	return reply, true
 }
@@ -753,6 +927,26 @@ func (h *reactiveHandler) queryKiwix(ctx context.Context, t *queryTurn) (string,
 		}
 	}
 
+	// The topic, not the sentence (V-668). Kiwix ranks by keyword overlap, so
+	// the question words outrank the one word that names the article: measured
+	// on 2026-08-09, "что такое TCP" returns "Перехват TCP-соединения" and
+	// "TCP" returns TCP. Only the verbatim path needs this. The rewriter
+	// already reduces a question to English keywords, and reducing twice would
+	// take the topic off the input it reads.
+	if verbatim {
+		if topic := kiwix.Topic(pattern); topic != "" {
+			// The article named exactly, before any ranking runs. A ZIM is
+			// addressable by title and a wrong title is a 404, so this either
+			// answers or costs one request that says nothing.
+			page, err := h.kiwix.client.Article(ctxK, kiwix.TitlePath(book, topic), h.kiwix.runes)
+			if err == nil && page.Text != "" {
+				log.Printf("voice: kiwix: %q in %q → title hit %q", topic, book, page.Title)
+				return h.kiwixReply(ctx, t, page.Title, page.Text)
+			}
+			pattern = topic
+		}
+	}
+
 	hits, err := h.kiwix.client.Search(ctxK, pattern, book, h.kiwix.max)
 	if err != nil {
 		log.Printf("voice: kiwix: search %q: %v", pattern, err)
@@ -785,14 +979,18 @@ func (h *reactiveHandler) queryKiwix(ctx context.Context, t *queryTurn) (string,
 		}
 		page = crawl.Page{Title: top.Title, Text: top.Snippet}
 	}
-	// Handed over the same way a note or a page is: context for the question he
-	// asked, not something to recite.
-	snippet := top.Title + "\n" + crawl.TrimRunes(page.Text, h.kiwix.runes)
+	return h.kiwixReply(ctx, t, top.Title, page.Text)
+}
+
+// kiwixReply hands one article over the same way a note or a page is handed
+// over: context for the question he asked, not something to recite.
+func (h *reactiveHandler) kiwixReply(ctx context.Context, t *queryTurn, title, text string) (string, bool) {
+	snippet := title + "\n" + crawl.TrimRunes(text, h.kiwix.runes)
 	reply := h.phraseSource(ctx, "kiwix", t.dec.Utterance, []string{snippet})
 	if reply == "" {
 		// No phraser, or it failed. Read back the best hit rather than pretend
 		// the search did not happen.
-		return phraser.Q(phraser.QueryFound, map[string]string{"text": crawl.TrimRunes(top.Title+" — "+page.Text, 300)}), true
+		return readBack(title + " — " + text), true
 	}
 	return reply, true
 }

cmd/mavend/actions_reminder.go

@@ -19,7 +19,10 @@ func (h *reactiveHandler) actionReminder(ctx context.Context, dec router.Decisio
 		// time wasn't parsed. Run the parser as a fallback.
 		if dec.Stage == 0 && h.timeParser != nil {
 			t, ok, err := h.timeParser.Parse(ctx, dec.Utterance, h.now())
-			if err == nil && ok {
+			// Same gate as the extractor (V-577, V-579, V-610): a request whose
+			// hour was not spoken, or was spoken and not read, gets asked about
+			// and is never completed from the clock.
+			if err == nil && ok && router.ResolvedTheHour(dec.Utterance, t) {
 				dec.Slots.Time = t
 				dec.Slots.HasTime = true
 			}

cmd/mavend/boot.go

@@ -0,0 +1,120 @@
+package main
+
+import (
+	"context"
+	"errors"
+	"log"
+	"net"
+	"sync"
+	"time"
+
+	"github.com/kami/maven/internal/event"
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/store"
+)
+
+// The two boot paths meet here. run() wires the daemon twice: once at boot
+// when a key is in the environment, and once inside UnlockFn after a passkey
+// assertion, minutes or days later. Listing the same wiring in both places is
+// what let them drift — seven workers started untracked on the unlock path and
+// two daemonAPI fields were never set there, silently, for as long as anyone
+// had been cold-starting (V-639).
+//
+// So both paths call newDaemonAPI and startBackground and nothing else. A
+// field or a worker added later reaches both paths or neither.
+
+// bootDeps is everything the two constructors below read. It is filled from
+// the same variables on both paths, by depsNow in run().
+type bootDeps struct {
+	coreFor    func() ipc.CoreAPI
+	tl         *tickLoop
+	evBus      *event.Bus
+	voiceW     *voiceWiring
+	st         *store.Store
+	factWorker *factEnrichmentWorker
+	evalWorker *memoryEvalWorker // nil ⇒ memory evaluation off (the default)
+	feedWkr    *feedWorker       // nil ⇒ no feed is read (the default)
+	crawlWkr   *crawlWorker      // nil ⇒ no page is watched (the default)
+}
+
+// newDaemonAPI builds the real CoreAPI, with every field set. The unlock path
+// used to leave nexus and getMCPServers nil, so after a cold start
+// ResolveEntity refused with a nexus block configured and /tools rendered
+// "not configured" with an mcp block configured. Empty is a wrong answer
+// there, not a degraded one.
+func newDaemonAPI(d bootDeps) *daemonAPI {
+	api := &daemonAPI{
+		CoreAPI:          d.coreFor(),
+		getTrace:         d.tl.trace,
+		getMorningStatus: func(ctx context.Context) []ipc.MorningRoutineStatus { return d.tl.morningStatus(ctx, time.Now()) },
+		getDayPlan:       func(ctx context.Context) ipc.DayPlan { return d.tl.dayPlan(ctx, time.Now()) },
+		getEvents:        intakeEventsFn(d.evBus),
+		getDecisions:     turnDecisionsFn(d.voiceW),
+		seedStore:        seedStoreIfAllowed(d.st),
+		nexus:            nexusOf(d.voiceW),
+	}
+	if d.voiceW != nil && d.voiceW.handler != nil {
+		api.chatFn = d.voiceW.handler.handleText
+		// And the reverse: the handler was wired with the bare store adapter,
+		// which cannot serve the day plan. See upgradeAPI.
+		d.voiceW.handler.upgradeAPI(api)
+	}
+	if d.voiceW != nil && d.voiceW.mcp != nil {
+		api.getMCPServers = d.voiceW.mcp.status
+	}
+	return api
+}
+
+// namedWorker is one long-running goroutine. The name exists so the set is
+// assertable from a test and readable in a log; nothing dispatches on it.
+type namedWorker struct {
+	name string
+	run  func(ctx context.Context)
+}
+
+// backgroundWorkers lists what this deployment runs. It is pure — it starts
+// nothing — so a test can compare the set the two paths would start without
+// standing a daemon up.
+func backgroundWorkers(d bootDeps) []namedWorker {
+	var ws []namedWorker
+	if d.voiceW != nil && d.voiceW.server != nil {
+		ws = append(ws, namedWorker{"voice", func(context.Context) {
+			if err := d.voiceW.server.Serve(); err != nil && !errors.Is(err, net.ErrClosed) {
+				log.Printf("voice serve: %v", err)
+			}
+		}})
+	}
+	ws = append(ws,
+		namedWorker{"tick", d.tl.run},
+		namedWorker{"fact-enrichment", d.factWorker.run},
+	)
+	if d.evalWorker != nil {
+		ws = append(ws, namedWorker{"memory-eval", d.evalWorker.run})
+	}
+	if d.feedWkr != nil {
+		ws = append(ws, namedWorker{"feed", d.feedWkr.run})
+	}
+	if d.crawlWkr != nil {
+		ws = append(ws, namedWorker{"crawl", d.crawlWkr.run})
+	}
+	if d.voiceW != nil && d.voiceW.mcp != nil {
+		ws = append(ws, namedWorker{"mcp", d.voiceW.mcp.run})
+	}
+	if d.voiceW != nil && d.voiceW.home != nil {
+		ws = append(ws, namedWorker{"home", d.voiceW.home.run})
+	}
+	return ws
+}
+
+// startBackground starts every worker through goWorker, so waitWorkers can
+// wait for it at shutdown. A worker started as a bare `go func()` is the
+// shutdown bug documented at the end of run(): run() never returns, the
+// deferred Close never seals the database, and the ciphertext goes stale.
+func startBackground(ctx context.Context, wg *sync.WaitGroup, d bootDeps) {
+	for _, w := range backgroundWorkers(d) {
+		goWorker(wg, func() { w.run(ctx) })
+	}
+	if d.voiceW != nil && d.voiceW.server != nil {
+		log.Printf("mavend: voice listening on %s", d.voiceW.server.Addr())
+	}
+}

cmd/mavend/boot_test.go

@@ -0,0 +1,96 @@
+package main
+
+import (
+	"reflect"
+	"testing"
+
+	"github.com/kami/maven/internal/decision"
+	"github.com/kami/maven/internal/event"
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/store"
+	"github.com/kami/maven/internal/voice"
+)
+
+// fullDeps — a deployment with every optional piece present. Nothing here is
+// run: newDaemonAPI takes method values and backgroundWorkers is pure, so
+// zero-value wirings are enough to say what WOULD be started.
+func fullDeps() bootDeps {
+	h := &reactiveHandler{
+		ecosystem: &ecosystemWiring{nexus: &nexusClient{}},
+		decisions: decision.NewRing(),
+	}
+	return bootDeps{
+		coreFor:    func() ipc.CoreAPI { return ipc.UnimplementedCoreAPI{} },
+		tl:         &tickLoop{},
+		evBus:      event.NewBus(4),
+		st:         &store.Store{},
+		factWorker: &factEnrichmentWorker{},
+		evalWorker: &memoryEvalWorker{},
+		feedWkr:    &feedWorker{},
+		crawlWkr:   &crawlWorker{},
+		voiceW: &voiceWiring{
+			server:  &voice.Server{},
+			handler: h,
+			mcp:     &mcpWiring{},
+			home:    &homeWiring{},
+		},
+	}
+}
+
+// The unlock path used to build its own daemonAPI literal and leave nexus and
+// getMCPServers nil (V-639). Both paths call newDaemonAPI now, so the drift
+// that can still happen is a field added to the struct and not to the
+// constructor. This catches that one, by name.
+func TestNewDaemonAPISetsEveryField(t *testing.T) {
+	prev := allowSeedOnStart
+	allowSeedOnStart = true
+	defer func() { allowSeedOnStart = prev }()
+
+	api := newDaemonAPI(fullDeps())
+	v := reflect.ValueOf(*api)
+	for i := range v.NumField() {
+		if v.Field(i).IsZero() {
+			t.Errorf("newDaemonAPI left %s unset — a fully wired deployment must fill every field", v.Type().Field(i).Name)
+		}
+	}
+}
+
+// The handler is wired with the bare store adapter and cannot serve the day
+// plan until upgradeAPI hands it the real one. The unlocked path did that and
+// the unlock path did it too; keep it a property of the constructor.
+func TestNewDaemonAPIUpgradesTheHandler(t *testing.T) {
+	d := fullDeps()
+	api := newDaemonAPI(d)
+	if d.voiceW.handler.api != ipc.CoreAPI(api) {
+		t.Fatal("newDaemonAPI did not hand the handler the API it built")
+	}
+}
+
+// Every worker the daemon runs goes through startBackground, so shutdown can
+// wait for it. The unlock path used to start seven of these as bare
+// `go func()` under a shadowed WaitGroup.
+func TestBackgroundWorkersFullSet(t *testing.T) {
+	want := []string{"voice", "tick", "fact-enrichment", "memory-eval", "feed", "crawl", "mcp", "home"}
+	var got []string
+	for _, w := range backgroundWorkers(fullDeps()) {
+		got = append(got, w.name)
+	}
+	if !reflect.DeepEqual(got, want) {
+		t.Errorf("workers = %v, want %v", got, want)
+	}
+}
+
+// A default box configures none of the optional blocks. Two workers always run
+// and the rest stay dark, rather than a nil run being scheduled.
+func TestBackgroundWorkersFloor(t *testing.T) {
+	d := fullDeps()
+	d.evalWorker, d.feedWkr, d.crawlWkr, d.voiceW = nil, nil, nil, nil
+	want := []string{"tick", "fact-enrichment"}
+	var got []string
+	for _, w := range backgroundWorkers(d) {
+		got = append(got, w.name)
+	}
+	if !reflect.DeepEqual(got, want) {
+		t.Errorf("workers = %v, want %v", got, want)
+	}
+}

cmd/mavend/chat_degrade_test.go

@@ -19,7 +19,7 @@ func TestChatAnswersWithNoLlamaServer(t *testing.T) {
 	dead := llm.New("http://127.0.0.1:1", 500*time.Millisecond)
 	emb := router.NewHashEmbedder(1024)
 	h.recall.embedder = emb
-	h.router = buildRouter(emb, h.matcher, 0.55, pickLLMRouter(true, dead))
+	h.router = buildRouter(emb, h.matcher, 0.55, pickLLMRouter(true, dead), nil)
 	h.replier = newLLMReplier(dead, nil)
 
 	ctx := withDialogueID(context.Background(), dialogueIDFor(sourceText, "web"))

cmd/mavend/clarify.go

@@ -6,6 +6,7 @@ import (
 	"math/rand"
 	"strings"
 	"time"
+	"unicode/utf8"
 
 	"github.com/kami/maven/internal/dialogue"
 	"github.com/kami/maven/internal/router"
@@ -65,8 +66,28 @@ var clarifyExpiredVariants = []string{
 	"Столько времени прошло, что я отпустила прошлую просьбу. Скажи заново, если она в силе.",
 }
 
-// clarifyExpiredLine picks one of them at random.
-func clarifyExpiredLine() string {
+// clarifyExpiredPluralVariants — the same notice when TWO parked requests died
+// together (Vikunja #561). Since a side query suspends the flow instead of
+// dropping it, the stack can hold both the flow and the thing he interrupted it
+// with, and TakeExpired drops the whole stack when the top times out. "Прошлую
+// просьбу" would then be a lie about the count: he loses two and hears about
+// one.
+//
+// Two phrasings only, against five for the singular. This fires when he walks
+// off in the middle of an interrupted exchange, which is rarer than walking off
+// in the middle of a plain one, so it repeats less and needs less variety.
+var clarifyExpiredPluralVariants = []string{
+	"Прости, я слишком долго ждала и отпустила обе прошлые просьбы. Если они ещё нужны, скажи заново.",
+	"Я не дождалась ответа и убрала обе прошлые просьбы. Повтори, если они всё ещё нужны.",
+}
+
+// clarifyExpiredLine picks one of them at random. n is how many requests died;
+// anything above one gets the plural wording, because the bound is two today and
+// a third would still be "обе" short of the truth only if MaxStackDepth grew.
+func clarifyExpiredLine(n int) string {
+	if n > 1 {
+		return clarifyExpiredPluralVariants[rand.Intn(len(clarifyExpiredPluralVariants))]
+	}
 	return clarifyExpiredVariants[rand.Intn(len(clarifyExpiredVariants))]
 }
 
@@ -74,23 +95,34 @@ func clarifyExpiredLine() string {
 // notice is glued in front of this turn's reply (see withNotice), so a caller
 // checking for it has to match a prefix, not the whole string.
 func isClarifyExpired(s string) bool {
-	for _, v := range clarifyExpiredVariants {
-		if strings.HasPrefix(s, v) {
-			return true
-		}
-	}
-	return false
+	_, ok := cutClarifyExpired(s)
+	return ok
 }
 
 // trimClarifyExpired strips a leading expiry notice, leaving this turn's actual
 // reply. "" ⇒ the notice was the whole thing.
 func trimClarifyExpired(s string) string {
-	for _, v := range clarifyExpiredVariants {
-		if strings.HasPrefix(s, v) {
-			return strings.TrimSpace(strings.TrimPrefix(s, v))
+	rest, ok := cutClarifyExpired(s)
+	if !ok {
+		return strings.TrimSpace(s)
+	}
+	return rest
+}
+
+// cutClarifyExpired matches either expiry deck as a prefix and returns what
+// follows it. Both decks, since V-561 added the plural line: a caller asking
+// "did she say a request timed out" means the fact, not which wording carried
+// it, and a helper that knew only the singular would read the two-request
+// notice as ordinary reply text.
+func cutClarifyExpired(s string) (string, bool) {
+	for _, deck := range [][]string{clarifyExpiredVariants, clarifyExpiredPluralVariants} {
+		for _, v := range deck {
+			if strings.HasPrefix(s, v) {
+				return strings.TrimSpace(strings.TrimPrefix(s, v)), true
+			}
 		}
 	}
-	return strings.TrimSpace(s)
+	return "", false
 }
 
 // clarifyExpiredNotice returns that line when a parked question had just timed
@@ -101,11 +133,12 @@ func (h *reactiveHandler) clarifyExpiredNotice(ctx context.Context) string {
 	if h.clarifyStore == nil {
 		return ""
 	}
-	if !h.clarifyStore.TakeExpired(dialogueIDOf(ctx), h.now()) {
+	n := h.clarifyStore.TakeExpired(dialogueIDOf(ctx), h.now())
+	if n == 0 {
 		return ""
 	}
-	log.Printf("voice: clarify — parked question expired, telling him and routing the words fresh")
-	return clarifyExpiredLine()
+	log.Printf("voice: clarify — %d parked question(s) expired, telling him and routing the words fresh", n)
+	return clarifyExpiredLine(n)
 }
 
 // withNotice glues the expiry notice in front of this turn's reply. One turn
@@ -121,10 +154,47 @@ func withNotice(notice, reply string) string {
 	return notice + " " + reply
 }
 
+// withResumed puts the resumed question on the END of this turn's reply, where
+// withNotice puts the expiry notice on the front (Vikunja #561).
+//
+// The order is the owner's: "в Риме сейчас ..., на какое время поставить
+// напоминание?" — answer first, then the open question. A question in front of
+// its own answer would read as ignoring what he asked.
+//
+// Two sentences, not one (V-654). This used to fold the answer's full stop into
+// a comma, on the strength of the owner having written it that way once. Spliced
+// onto a real answer it reads as one run-on thought — "вот что я нашла: вайфай
+// пароль лежит в ящике стола, на какое время поставить напоминание?" — and the
+// question disappears into the tail of a sentence about something else. A reply
+// with no terminator of its own is given one, so the join never depends on how
+// the phraser chose to end.
+//
+// A resume with no answer in front of it is just the question.
+func withResumed(reply, resumed string) string {
+	if resumed == "" {
+		return reply
+	}
+	reply = strings.TrimSpace(reply)
+	if reply == "" {
+		return resumed
+	}
+	if !endsSentence(reply) {
+		reply += "."
+	}
+	return reply + " " + resumed
+}
+
+// endsSentence reports whether s already closes itself. The ellipsis counts: a
+// trailing "…" is a deliberate end, and a full stop after it reads as a typo.
+func endsSentence(s string) bool {
+	r, _ := utf8.DecodeLastRuneInString(s)
+	return strings.ContainsRune(".!?…", r)
+}
+
 // missingFor returns the slots a decision still needs, most important first.
 // Empty ⇒ there is nothing identifiable to ask about.
 func missingFor(dec router.Decision) []dialogue.Slot {
-	return dialogue.StillMissing(wantedSlots[dec.Intent], toDialogueSlots(dec.Slots))
+	return stillMissingFor(dec.Intent, dec.Utterance, toDialogueSlots(dec.Slots))
 }
 
 // clarifyQuestion picks the one question to ask for a clarify decision. Returns
@@ -133,18 +203,32 @@ func missingFor(dec router.Decision) []dialogue.Slot {
 // One question about one thing: if two slots are missing she asks about the
 // first only. Two questions in one breath is an interrogation. The second gap
 // is picked up on the turn after the first one is answered (askRemainingGap).
-func clarifyQuestion(dec router.Decision) (dialogue.Slot, string, bool) {
+func (h *reactiveHandler) clarifyQuestion(dec router.Decision) (dialogue.Slot, string, bool) {
 	missing := missingFor(dec)
 	if len(missing) == 0 {
 		return "", "", false
 	}
-	q, ok := clarifyQuestionFor(missing[0], 1)
+	q, ok := h.questionFor(missing[0], 1, dec.Utterance, toDialogueSlots(dec.Slots), "")
 	if !ok {
 		return "", "", false
 	}
 	return missing[0], q, true
 }
 
+// questionFor picks the wording for one gap. Every slot but the reminder's time
+// reads its deck by attempt; the time asks about whichever of the hour, the half
+// of the day and the day he has not said, and states the clock while it does
+// (V-579).
+//
+// taken is the acknowledgement of what his last turn added, empty when it added
+// nothing and empty for a first ask, which has no turn behind it (V-593).
+func (h *reactiveHandler) questionFor(slot dialogue.Slot, attempt int, utterance string, s dialogue.Slots, taken string) (string, bool) {
+	if slot != dialogue.SlotTime {
+		return clarifyQuestionFor(slot, attempt)
+	}
+	return whenQuestion(whenGapOf(utterance, s.HasTime), attempt, h.now(), taken)
+}
+
 // askClarify parks the request and returns the question to ask instead of the
 // canned "не поняла". Returns ("", false) when there is nothing to ask about, so
 // the caller falls back to the canned reply.
@@ -152,11 +236,16 @@ func (h *reactiveHandler) askClarify(ctx context.Context, dec router.Decision) (
 	if h.clarifyStore == nil {
 		return "", false
 	}
-	slot, question, ok := clarifyQuestion(dec)
+	slot, question, ok := h.clarifyQuestion(dec)
 	if !ok {
 		return "", false
 	}
-	h.clarifyStore.Put(dialogueIDOf(ctx), &dialogue.PendingQuestion{
+	// An act she could not resolve is a refusal, not a question (Vikunja #556).
+	if slot == dialogue.SlotFn {
+		log.Printf("voice: clarify — act %q matched no capability; saying so instead of asking", dec.Utterance)
+		return actNotRecognized, true
+	}
+	q := &dialogue.PendingQuestion{
 		Intent:      dialogue.Intent(dec.Intent),
 		Slots:       toDialogueSlots(dec.Slots),
 		Missing:     []dialogue.Slot{slot},
@@ -165,70 +254,170 @@ func (h *reactiveHandler) askClarify(ctx context.Context, dec router.Decision) (
 		TTL:         clarifyTTL,
 		Attempts:    1, // this ask
 		MaxAttempts: h.clarifyMaxAttempts,
-	})
+	}
+	// Push, not Put, when this turn suspended a flow (Vikunja #561): the side
+	// query needs clarifying of ITS own, and Put would replace the top of the
+	// stack — which is the very question the side query was allowed to interrupt
+	// rather than kill. Push keeps both.
+	//
+	// Push returns whatever the depth bound forced out, and that one has to be
+	// spoken: MaxStackDepth is a promise that every level she keeps is a level
+	// she can name when it dies. It is glued in front, like every other notice
+	// about something let go.
+	rt := turnRouteFrom(ctx)
+	if rt != nil && rt.suspended {
+		if evicted := h.clarifyStore.Push(dialogueIDOf(ctx), q); evicted != nil {
+			log.Printf("voice: clarify — stack full at %d, letting go of the request behind %q", dialogue.MaxStackDepth, evicted.Utterance)
+			rt.dropped = withNotice(rt.dropped, clarifyDropped)
+		}
+		// One question per breath still holds. The side query turned out to need
+		// a question of its own, so THAT is the one she asks; resuming as well
+		// would put two questions in one reply, which is the interrogation
+		// askRemainingGap already refuses to run. The suspended flow keeps its
+		// place underneath and is not lost — if it is never reached it dies on
+		// the TTL, and the expiry notice (now plural-aware) says so.
+		rt.resume = ""
+		log.Printf("voice: clarify — asked about %s for intent=%s, stacked on a suspended flow", slot, dec.Intent)
+		return question, true
+	}
+	h.clarifyStore.Put(dialogueIDOf(ctx), q)
 	log.Printf("voice: clarify — asked about %s for intent=%s", slot, dec.Intent)
 	return question, true
 }
 
-// resolveClarifyAnswer reads an utterance as the answer to a parked question.
-// Returns ("", false) when no live question is parked (or it expired), so the
-// caller routes the utterance normally as a fresh request. Sibling of
-// resolveConfirm and checked in the same place.
-//
-// The answer is parsed with the same extractor the router uses, for the intent
-// she parked — no second parser. If it still does not fill the gap she asks
-// again, up to MaxAttempts; after that she says out loud that she did not
-// understand. She never drops the request in silence.
-// isOwnRequest reports whether an utterance asks for something in its own
-// right, which is what a clarify answer never does. Two offline tests over
-// tokens, both already written for other callers: a question shape, and a
-// capture verb. Cheap on purpose — this runs on the answer to every parked
-// question, and it must not cost a model call.
-//
-// It is not a general relevance test. A bare noun that answers nothing ("синий"
-// after "Что сделать?") is still treated as an answer and still re-asked, and
-// that is the intended shape: only an utterance that carries its own request
-// wins over the question in front of it.
-func isOwnRequest(text string) bool {
-	return router.IsQuestionShaped(text) || router.CarriesCaptureVerb(text)
-}
+// clarifyCancelled — he called the half-built request off. Said out loud, like
+// every other way it can end: a silent drop reads as "done". Feminine
+// self-reference ("отменила"), as everywhere.
+const clarifyCancelled = "Хорошо, отменила."
 
+// clarifyDropped — he asked for something ELSE instead, so the parked request
+// is gone. Glued in front of the answer to what he actually asked, because
+// nothing may be dropped in silence.
+//
+// Only new_request and cancel reach this since V-561. A side query used to as
+// well, and the owner rejected it on sight: he asks about the weather in the
+// middle of setting a reminder, and hearing "прошлую просьбу отпускаю" tells him
+// a thing he did not ask to lose has been lost. It had not been — there was
+// simply nowhere to put it. Now there is (ClarifyStore's stack), so a side query
+// suspends and resumes, and apologising for a drop that did not happen is worse
+// than saying nothing.
+const clarifyDropped = "Прошлую просьбу отпускаю."
+
+// resolveClarifyAnswer reads an utterance against the parked question and
+// decides what it IS before deciding what to do with it. Returns ("", false)
+// when the turn is not this resolver's — nothing parked, or the utterance turned
+// out to be a request of its own — so the caller dispatches it normally.
+//
+// The order is the point (Vikunja #560). The utterance is ROUTED first, and the
+// role is read off that decision: a routed decision that stands on its own is
+// not an answer, whatever the extractor found inside it. Before this the
+// extractor decided, so "какая сейчас погода в Риме?" became the time of a
+// reminder on the strength of the word "сейчас".
+//
+// The answer itself is parsed with the same extractor the router uses, for the
+// intent she parked — no second parser. If it still does not fill the gap she
+// asks again, up to MaxAttempts; after that she says out loud that she did not
+// understand. She never drops the request in silence.
 func (h *reactiveHandler) resolveClarifyAnswer(ctx context.Context, text string) (string, bool) {
 	if h.clarifyStore == nil {
 		return "", false
 	}
 	q := h.clarifyStore.Get(dialogueIDOf(ctx), h.now())
 	if q == nil {
-		return "", false
+		return "", false // not_applicable: nothing is pending
 	}
 
 	intent := router.Intent(q.Intent)
 	answer := h.extractor.Extract(ctx, intent, text, h.now())
-	merged := q.Answer(text, toDialogueSlots(answer))
-	// He moved on. A parked question used to swallow whatever came next, so one
-	// act she could not fulfil ate the following three turns: "выключи свет в
-	// спальне" asked "Что сделать?", and "кто изобрёл телефон" was scored as an
-	// answer to it, then "как дела" after that (Vikunja #554). Nothing checked
-	// whether the words could be an answer at all.
-	//
-	// Deliberately narrow. It only fires where the answer filled nothing, so a
-	// turn that closes the gap is still an answer whatever shape it has, and
-	// the retry budget is untouched — the count was never the problem. Dropping
-	// the question and routing the utterance as itself is what he meant either
-	// way: if he really was answering, he can say it again, and if he was not,
-	// he gets the thing he asked for instead of being asked a third time.
-	if len(dialogue.StillMissing(q.Missing, merged)) > 0 && isOwnRequest(text) {
+
+	var (
+		routed   router.Decision
+		routedOK bool
+	)
+	if needsRoute(text) {
+		routed, routedOK = h.routeForRole(ctx, text)
+	}
+	role := classifyTurnRole(q, text, toDialogueSlots(answer), routed, routedOK)
+	log.Printf("voice: clarify — %q is a %s against %s (routed=%v)", text, role, dialogue.CapabilityFor(q.Intent), routedOK)
+
+	switch role {
+	case roleCancel:
 		h.clarifyStore.Delete(dialogueIDOf(ctx))
-		log.Printf("voice: clarify — %q is its own request, not an answer to %v; dropping the question", text, q.Missing)
+		return clarifyCancelled, true
+	case roleSideQuery:
+		// He asked something of his own WITHOUT leaving the flow. The question
+		// stays exactly where it is — same slot, same attempt, same parked
+		// utterance — and these words go on to be answered as themselves. The
+		// resumed question is then glued onto the back of that answer, so one
+		// reply carries both acts (Vikunja #561).
+		//
+		// No attempt is spent. He answered the side query, not the parked
+		// question, and charging a retry for a turn that was never an answer is
+		// the V-554 shape.
+		h.noteSuspended(ctx, q)
+		return "", false
+	case roleAside:
+		// He stated something in the middle of the flow. Same machinery as a
+		// side query and for the same reason: the words are answered as
+		// themselves, so the note or the fact is stored, and the question comes
+		// back on the end of the same reply (V-577 shape 2). Storing it in
+		// silence and dropping it in silence are both wrong, and dropping it is
+		// what she did.
+		h.noteSuspended(ctx, q)
+		return "", false
+	case roleNewRequest:
+		// He moved on. A parked question used to swallow whatever came next, so
+		// one act she could not fulfil ate the following three turns (Vikunja
+		// #554) and a world question set a reminder for a time nobody asked for
+		// (#558). Drop the question, say so, and let these words be themselves.
+		h.clarifyStore.Delete(dialogueIDOf(ctx))
+		h.noteDropped(ctx)
 		return "", false
 	}
+
+	// He is answering, so the run of step-asides is over (V-654). Reset here
+	// rather than where a gap is FILLED: "позвонить маме" against a question
+	// about the time gives her nothing she asked for and still means he is in
+	// the exchange, and the retry it costs is bound enough on its own. The
+	// counter is for the case the bounds miss — he asked for other things and
+	// never came back.
+	q.Suspends = 0
+
+	merged := q.Answer(text, toDialogueSlots(answer))
 	// Fold a newly answered subject into the raw utterance. Downstream actions
 	// phrase from Utterance, not from the text slot — actionReminder stores it
 	// as the reminder payload — so a reminder clarified out of a bare "напомни"
 	// would fire at 11:00 saying "напомни" and nothing else.
 	q.Utterance = foldAnswerIntoUtterance(q.Utterance, merged.Text)
-	if len(dialogue.StillMissing(q.Missing, merged)) > 0 {
-		return h.reaskOrGiveUp(ctx, q, merged, text), true
+	// A fact answers with a key and a value and fills no Text slot at all, so
+	// the fold above leaves the utterance at the bare "запиши" — and that is
+	// what the confirmation now reads back to him (V-592). His raw words are the
+	// only record of what he said, so they are what is folded. Never for a time
+	// question: what he says about when is kept apart in WhenText on purpose,
+	// or the reminder would read the day back at him when it fires.
+	if merged.Text == "" && !asksAboutTime(q.Missing) {
+		q.Utterance = foldAnswerIntoUtterance(q.Utterance, text)
+	}
+	// An answer about the time joins everything else he has said about the time,
+	// and the whole of it is re-read as one request (V-579). "завтра" names the
+	// day of an hour she is already holding, and read alone it names no hour at
+	// all, so the parser would have nothing and she would ask for ever.
+	// What he had already said about the time, read BEFORE this answer joins it.
+	// A re-ask that cannot tell the two apart is the one that repeats itself
+	// byte for byte (V-593).
+	var taken string
+	if asksAboutTime(q.Missing) {
+		before := whenKnownOf(whenTextOf(q), q.Slots.HasTime)
+		q.WhenText = strings.TrimSpace(q.WhenText + " " + text)
+		if t, ok := h.readWhen(ctx, intent, q, text); ok {
+			merged.Time, merged.HasTime = t, true
+		}
+		if before.movedForward(whenKnownOf(whenTextOf(q), merged.HasTime)) {
+			taken = whenTakenLine(text)
+		}
+	}
+	if stillOpen(q.Missing, whenTextOf(q), merged) {
+		return h.reaskOrGiveUp(ctx, q, merged, text, taken), true
 	}
 	h.clarifyStore.Delete(dialogueIDOf(ctx))
 
@@ -257,6 +446,64 @@ func (h *reactiveHandler) resolveClarifyAnswer(ctx context.Context, text string)
 	return h.finishClarified(ctx, dec), true
 }
 
+// noteDropped records that the parked request was let go this turn, so runTurn
+// can say it in front of whatever these words are answered with. Nothing to
+// record outside runTurn — a unit test calling one resolver has no turn to glue
+// a notice onto.
+func (h *reactiveHandler) noteDropped(ctx context.Context) {
+	if rt := turnRouteFrom(ctx); rt != nil {
+		rt.dropped = clarifyDropped
+	}
+}
+
+// noteSuspended keeps the parked question alive across a side query and records
+// the words that bring it back, so runTurn can put them after this turn's answer
+// (Vikunja #561).
+//
+// Two things happen to the question and neither is an attempt. Its clock is
+// restarted, because she is about to ask it again and the 90s TTL measures the
+// pause since she last spoke it — leaving Asked at the original ask would let a
+// flow he is actively working through die of a wait he did not take. And the
+// stack is left exactly as it is: the question is already on top, so suspending
+// it is not a write.
+//
+// A slot with no resumed wording (clarifyResumedFor says so) resumes nothing and
+// says nothing. She must not claim to be holding a question she cannot re-ask.
+//
+// Suspension is bounded, since V-654. Neither of the two things above is a
+// limit: no attempt is spent, and restarting the clock means the TTL cannot
+// arrive while he keeps talking. So the count is the only thing that ends it,
+// and past MaxSuspends she lets the request go and says so with the same line
+// every other drop uses. The rule is unchanged — a question ends by being
+// answered or by being let go out loud — this only recognises three unrelated
+// requests in a row as the second of those.
+func (h *reactiveHandler) noteSuspended(ctx context.Context, q *dialogue.PendingQuestion) {
+	rt := turnRouteFrom(ctx)
+	if rt == nil || len(q.Missing) == 0 {
+		return
+	}
+	question, ok := clarifyResumedFor(q.Missing[0])
+	if !ok {
+		return
+	}
+	if !q.CanResume() {
+		h.clarifyStore.Delete(dialogueIDOf(ctx))
+		h.noteDropped(ctx)
+		log.Printf("voice: clarify — letting the question about %s go: %d asides in a row, %d rides in all", q.Missing[0], q.Suspends, q.Rides)
+		return
+	}
+	q.Suspends++
+	// Rides is the same event counted without the reset (V-663). Incremented
+	// beside Suspends and never anywhere else, so the two cannot disagree about
+	// what happened, only about how much of it they remember.
+	q.Rides++
+	q.Asked = h.now()
+	h.clarifyStore.Put(dialogueIDOf(ctx), q)
+	rt.resume = question
+	rt.suspended = true
+	log.Printf("voice: clarify — is its own request; suspending the question about %s and resuming it in the same reply (suspend %d of %d, ride %d of %d)", q.Missing[0], q.Suspends, dialogue.MaxSuspends, q.Rides, dialogue.MaxRides)
+}
+
 // foldAnswerIntoUtterance appends an answered subject to the original words,
 // unless they already carry it. "напомни" + "позвонить маме" reads as the
 // request he would have made in one breath. Nothing is appended when the
@@ -284,25 +531,35 @@ func foldAnswerIntoUtterance(utterance, subject string) string {
 // costs a question exactly like a second try at the first one does, so the cap
 // still bounds how many times she can speak before acting or letting go.
 func (h *reactiveHandler) askRemainingGap(ctx context.Context, q *dialogue.PendingQuestion, intent router.Intent, merged dialogue.Slots) (string, bool) {
-	remaining := dialogue.StillMissing(wantedSlots[intent], merged)
+	remaining := stillMissingFor(intent, whenTextOf(q), merged)
 	if len(remaining) == 0 {
 		return "", false
 	}
 	// Attempts+1 is the question she is about to ask, and the budget is shared
 	// with the re-ask path, so the second gap is worded like a second try.
-	question, ok := clarifyQuestionFor(remaining[0], q.Attempts+1)
+	question, ok := h.questionFor(remaining[0], q.Attempts+1, whenTextOf(q), merged, "")
 	if !ok || !q.CanAsk() {
 		return "", false
 	}
+	// Suspends is not carried, and by this point it is already zero: the answer
+	// path resets it (V-654). Left off the literal so the zero is stated where
+	// the struct is built, rather than inherited from a field nobody names.
+	//
+	// Rides IS carried, and that is the whole point of it (V-663). This is the
+	// same request under a second question, not a new one, so the turns it has
+	// already ridden still count against it. Dropping the field here is exactly
+	// the re-basing that let one question ride twenty-six replies.
 	h.clarifyStore.Put(dialogueIDOf(ctx), &dialogue.PendingQuestion{
 		Intent:      q.Intent,
 		Slots:       merged,
 		Missing:     []dialogue.Slot{remaining[0]},
 		Utterance:   q.Utterance,
+		WhenText:    q.WhenText,
 		Asked:       h.now(),
 		TTL:         clarifyTTL,
 		Attempts:    q.Attempts + 1,
 		MaxAttempts: q.MaxAttempts,
+		Rides:       q.Rides,
 	})
 	log.Printf("voice: clarify — one gap filled, still missing %s for intent=%s, asking again (attempt %d)", remaining[0], intent, q.Attempts+1)
 	return question, true
@@ -311,10 +568,12 @@ func (h *reactiveHandler) askRemainingGap(ctx context.Context, q *dialogue.Pendi
 // reaskOrGiveUp handles an answer that left the gap open: ask the same question
 // again while she has attempts left, otherwise say she did not understand and
 // let the request go. Never returns "" — a mute give-up reads as "done".
-func (h *reactiveHandler) reaskOrGiveUp(ctx context.Context, q *dialogue.PendingQuestion, merged dialogue.Slots, text string) string {
+// taken is the acknowledgement of what this answer DID give, empty when it gave
+// nothing (V-593). The give-up line never carries it: it is not another ask.
+func (h *reactiveHandler) reaskOrGiveUp(ctx context.Context, q *dialogue.PendingQuestion, merged dialogue.Slots, text, taken string) string {
 	question := ""
 	if len(q.Missing) > 0 {
-		question, _ = clarifyQuestionFor(q.Missing[0], q.Attempts+1)
+		question, _ = h.questionFor(q.Missing[0], q.Attempts+1, whenTextOf(q), merged, taken)
 	}
 	if question == "" || !q.CanAsk() {
 		h.clarifyStore.Delete(dialogueIDOf(ctx))
@@ -342,7 +601,7 @@ func (h *reactiveHandler) finishClarified(ctx context.Context, dec router.Decisi
 	}
 	reply := h.applyAction(ctx, dec)
 	if reply == "" {
-		reply = h.replier.Reply(dec)
+		reply = h.replier.Reply(ctx, dec)
 	}
 	if reply == "" {
 		// Belt: an empty reply here would be a silent drop.
@@ -351,20 +610,22 @@ func (h *reactiveHandler) finishClarified(ctx context.Context, dec router.Decisi
 	return reply
 }
 
+// maxCarriedHistory — how many turns of PRIOR history (beyond the immediate
+// last turn) rememberTurn carries forward. The session ends up holding this
+// many plus the one just-finished turn, so callers describing the total
+// depth (chatHistory's doc comment, this one) say "up to 4".
+const maxCarriedHistory = 3
+
 // rememberTurn stores this turn as the dialogue session the next follow-up
 // inherits from, carrying up to 4 prior turns of history for anaphora. Capped so
 // one long conversation can't grow the session unboundedly.
 func (h *reactiveHandler) rememberTurn(ctx context.Context, prev *dialogue.Session, dec router.Decision, now time.Time) {
 	var history []dialogue.Turn
 	if prev != nil {
-		history = append(history, dialogue.Turn{
-			Intent: prev.Intent,
-			Slots:  prev.Slots,
-			Text:   prev.Slots.Text,
-		})
+		history = append(history, sessionAsTurn(prev))
 		maxHist := len(prev.History)
-		if maxHist > 3 {
-			maxHist = 3
+		if maxHist > maxCarriedHistory {
+			maxHist = maxCarriedHistory
 		}
 		history = append(history, prev.History[:maxHist]...)
 	}

cmd/mavend/clarify_test.go

@@ -10,6 +10,7 @@ import (
 
 	"github.com/kami/maven/internal/dialogue"
 	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/memory"
 	"github.com/kami/maven/internal/phraser/eval"
 	"github.com/kami/maven/internal/router"
 	"github.com/kami/maven/internal/store"
@@ -57,18 +58,26 @@ func TestClarifyQuestionForMissingSlot(t *testing.T) {
 		want  string
 		asked bool
 	}{
-		{"reminder without a time", clarifyDec(router.IntentReminder, router.Slots{Text: "напомни позвонить маме"}, "напомни позвонить маме"), "Когда?", true},
+		{"reminder without a time", clarifyDec(router.IntentReminder, router.Slots{Text: "напомни позвонить маме"}, "напомни позвонить маме"), "Сейчас 09:00. Когда?", true},
 		{"fact without a key", clarifyDec(router.IntentFact, router.Slots{Text: "запиши"}, "запиши"), "Что записать?", true},
 		{"act without a fn", clarifyDec(router.IntentAct, router.Slots{Text: "сделай это"}, "сделай это"), "Что сделать?", true},
 		// A time with nothing to say at that time is still half a reminder, so
 		// the subject is what she asks about — not silence.
 		{"reminder that has a time but no subject", clarifyDec(router.IntentReminder, router.Slots{HasTime: true}, "напомни в 11"), "О чём напомнить?", true},
-		{"reminder that has both", clarifyDec(router.IntentReminder, router.Slots{Text: "позвонить маме", HasTime: true}, "напомни в 11 позвонить маме"), "", false},
+		// A bare hour is half of a day away from being an answer, and she asks
+		// which half rather than picking one (V-579).
+		{"reminder whose hour could be either half of the day", clarifyDec(router.IntentReminder, router.Slots{Text: "позвонить маме", HasTime: true}, "напомни в 11 позвонить маме"), "Сейчас 09:00. Это утра или вечера?", true},
+		{"reminder that has all three", clarifyDec(router.IntentReminder, router.Slots{Text: "позвонить маме", HasTime: true}, "напомни завтра в 15:00 позвонить маме"), "", false},
+		// The owner's own two, confirmed 2026-08-06: an unambiguous time and a
+		// relative one are both complete and are never asked about.
+		{"an interval names the instant by itself", clarifyDec(router.IntentReminder, router.Slots{Text: "позвонить маме", HasTime: true}, "напомни через час позвонить маме"), "", false},
+		{"half an hour is an interval too", clarifyDec(router.IntentReminder, router.Slots{Text: "выключить духовку", HasTime: true}, "напомни через полчаса выключить духовку"), "", false},
 		{"chat is never worth a question", clarifyDec(router.IntentChat, router.Slots{Text: "мгм"}, "мгм"), "", false},
 		{"query is never worth a question", clarifyDec(router.IntentQuery, router.Slots{Text: "а"}, "а"), "", false},
 	}
+	h, _, _ := newClarifyHandler(t)
 	for _, tc := range cases {
-		_, got, asked := clarifyQuestion(tc.dec)
+		_, got, asked := h.clarifyQuestion(tc.dec)
 		if asked != tc.asked || got != tc.want {
 			t.Errorf("%s: got (%q, %v), want (%q, %v)", tc.name, got, asked, tc.want, tc.asked)
 		}
@@ -82,11 +91,13 @@ func TestClarifyReminderCompletesOnAnswer(t *testing.T) {
 	h, st, _ := newClarifyHandler(t)
 
 	question, asked := h.askClarify(ctx, clarifyDec(router.IntentReminder, router.Slots{Text: "напомни позвонить маме"}, "напомни позвонить маме"))
-	if !asked || question != "Когда?" {
+	if !asked || question != "Сейчас 09:00. Когда?" {
 		t.Fatalf("expected the time question, got %q asked=%v", question, asked)
 	}
 
-	reply, handled := h.resolveClarifyAnswer(ctx, "в 11:00")
+	// The answer names the day as well as the hour. A reminder commits on what,
+	// what time and what day, and a dayless hour is asked about (V-579).
+	reply, handled := h.resolveClarifyAnswer(ctx, "сегодня в 11:00")
 	if !handled {
 		t.Fatal("the answer to an open question must be consumed as an answer")
 	}
@@ -158,11 +169,11 @@ func TestClarifyAsksThreeTimesThenSaysSo(t *testing.T) {
 		}
 		// The wording changes with the attempt (Vikunja #457): repeating a
 		// question he already failed to answer is the worst way to ask it.
-		want, _ := clarifyQuestionFor(dialogue.SlotTime, i)
+		want, _ := whenQuestion(whenNoHour, i, h.now(), "")
 		if reply != want {
 			t.Fatalf("attempt %d should ask again as %q, got %q", i, want, reply)
 		}
-		if first, _ := clarifyQuestionFor(dialogue.SlotTime, 1); reply == first {
+		if first, _ := whenQuestion(whenNoHour, 1, h.now(), ""); reply == first {
 			t.Fatalf("attempt %d repeated the first wording: %q", i, reply)
 		}
 		if h.clarifyStore.Get(voiceDialogueID, h.now()) == nil {
@@ -215,10 +226,11 @@ func TestClarifyRestatedAnswerWins(t *testing.T) {
 	if q == nil {
 		t.Fatal("expected an armed question")
 	}
-	first := h.extractor.Extract(ctx, router.IntentReminder, "в 11:00", h.now())
-	q.Slots = q.Answer("в 11:00", toDialogueSlots(first))
+	first := h.extractor.Extract(ctx, router.IntentReminder, "сегодня в 11:00", h.now())
+	q.Slots = q.Answer("сегодня в 11:00", toDialogueSlots(first))
+	q.WhenText = "сегодня в 11:00"
 
-	if reply, handled := h.resolveClarifyAnswer(ctx, "нет, в 15:00"); !handled || reply == clarifyGaveUp {
+	if reply, handled := h.resolveClarifyAnswer(ctx, "нет, сегодня в 15:00"); !handled || reply == clarifyGaveUp {
 		t.Fatalf("the restated answer should complete the request, handled=%v reply=%q", handled, reply)
 	}
 	reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour))
@@ -231,34 +243,35 @@ func TestClarifyRestatedAnswerWins(t *testing.T) {
 	}
 }
 
-// TestClarifiedActOffAllowlistIsStillRefused — clarification fills in an
-// argument, it never grants authority.
-func TestClarifiedActOffAllowlistIsStillRefused(t *testing.T) {
+// TestActOffAllowlistIsStillRefused — naming a capability is not being granted
+// one. Since Vikunja #556 an unresolved act no longer parks a question, so this
+// goes through applyAction, which is the only way an act runs.
+func TestActOffAllowlistIsStillRefused(t *testing.T) {
 	ctx := context.Background()
 	h, st, _ := newClarifyHandler(t)
 	marker := filepath.Join(t.TempDir(), "not-allowed-ran")
+	if err := st.EnableTool(ctx, "uptime", []string{"true"}, false, "test", h.now()); err != nil {
+		t.Fatal(err)
+	}
 
-	if _, asked := h.askClarify(ctx, clarifyDec(router.IntentAct, router.Slots{Text: "сделай это"}, "сделай это")); !asked {
-		t.Fatal("an act with no fn should be asked about")
-	}
-	reply, handled := h.resolveClarifyAnswer(ctx, "rm "+marker)
-	if !handled {
-		t.Fatal("the answer should be consumed")
-	}
+	reply := h.applyAction(ctx, router.Decision{
+		Utterance: "rm " + marker,
+		Intent:    router.IntentAct,
+		Slots:     router.Slots{Fn: "rm " + marker, HasFn: true},
+	})
 	if strings.Contains(reply, "готово") {
 		t.Fatalf("an act that is not on the allowlist must not report success: %q", reply)
 	}
 	if _, err := os.Stat(marker); !os.IsNotExist(err) {
 		t.Fatalf("a clarified act off the allowlist ran anyway: %v", err)
 	}
-	if tools, err := st.ListTools(ctx, "enabled"); err != nil || len(tools) != 0 {
-		t.Fatalf("clarify must not enable a tool: tools=%+v err=%v", tools, err)
+	if tools, err := st.ListTools(ctx, "enabled"); err != nil || len(tools) != 1 {
+		t.Fatalf("an act must not enable a tool: tools=%+v err=%v", tools, err)
 	}
 }
 
-// TestClarifiedDestructiveActStillNeedsConfirm — the confirm gate survives the
-// clarify path.
-func TestClarifiedDestructiveActStillNeedsConfirm(t *testing.T) {
+// TestDestructiveActStillNeedsConfirm — the confirm gate stands on the act path.
+func TestDestructiveActStillNeedsConfirm(t *testing.T) {
 	ctx := context.Background()
 	h, st, _ := newClarifyHandler(t)
 	marker := filepath.Join(t.TempDir(), "destructive-ran")
@@ -266,18 +279,16 @@ func TestClarifiedDestructiveActStillNeedsConfirm(t *testing.T) {
 		t.Fatal(err)
 	}
 
-	if _, asked := h.askClarify(ctx, clarifyDec(router.IntentAct, router.Slots{Text: "сделай это"}, "сделай это")); !asked {
-		t.Fatal("expected a question")
-	}
-	reply, handled := h.resolveClarifyAnswer(ctx, "delete_backups")
-	if !handled {
-		t.Fatal("the answer should be consumed")
-	}
+	reply := h.applyAction(ctx, router.Decision{
+		Utterance: "delete_backups",
+		Intent:    router.IntentAct,
+		Slots:     router.Slots{Fn: "delete_backups", HasFn: true},
+	})
 	if !strings.Contains(reply, "да") || h.pending == nil {
-		t.Fatalf("a clarified destructive act must still park a confirm: reply=%q pending=%+v", reply, h.pending)
+		t.Fatalf("a destructive act must park a confirm: reply=%q pending=%+v", reply, h.pending)
 	}
 	if _, err := os.Stat(marker); !os.IsNotExist(err) {
-		t.Fatalf("a clarified destructive act ran before confirmation: %v", err)
+		t.Fatalf("a destructive act ran before confirmation: %v", err)
 	}
 }
 
@@ -306,7 +317,7 @@ func TestClarifyExpiryIsAnnouncedAndWordsStillRoute(t *testing.T) {
 	h, _, now := newClarifyHandler(t)
 	emb := router.NewHashEmbedder(1024)
 	h.recall.embedder = emb
-	h.router = buildRouter(emb, h.matcher, 0.55, nil)
+	h.router = buildRouter(emb, h.matcher, 0.55, nil, nil)
 
 	if _, asked := h.askClarify(ctx, clarifyDec(router.IntentReminder, router.Slots{Text: "напомни"}, "напомни")); !asked {
 		t.Fatal("expected a question")
@@ -357,7 +368,7 @@ func TestClarifyAsksAboutTheSecondGapToo(t *testing.T) {
 	}
 	// Second gap, second attempt, so it is the second wording of the time
 	// question — the attempt budget is shared between the two paths.
-	want, _ := clarifyQuestionFor(dialogue.SlotTime, 2)
+	want, _ := whenQuestion(whenNoHour, 2, h.now(), "")
 	if reply != want {
 		t.Fatalf("a filled subject with no time must ask about the time as %q, got %q", want, reply)
 	}
@@ -369,7 +380,7 @@ func TestClarifyAsksAboutTheSecondGapToo(t *testing.T) {
 		t.Fatalf("the re-parked question lost the answered subject: %+v", q.Slots)
 	}
 
-	if reply, handled := h.resolveClarifyAnswer(ctx, "в 11:00"); !handled || reply == clarifyGaveUp {
+	if reply, handled := h.resolveClarifyAnswer(ctx, "сегодня в 11:00"); !handled || reply == clarifyGaveUp {
 		t.Fatalf("the time answer must complete the reminder, handled=%v reply=%q", handled, reply)
 	}
 	reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour))
@@ -419,7 +430,7 @@ func TestClarifyProseHoldsThePersona(t *testing.T) {
 		eval.CheckAddress:   true,
 		eval.CheckCringe:    true,
 	}
-	lines := append([]string{clarifyGaveUp}, clarifyExpiredVariants...)
+	lines := append([]string{clarifyGaveUp, clarifyCancelled, clarifyDropped}, clarifyExpiredVariants...)
 	lines = append(lines, clarifyMissedVariants...)
 	for _, variants := range clarifyQuestionVariants {
 		lines = append(lines, variants...)
@@ -487,7 +498,7 @@ func TestClarifySubjectAnswerFillsRatherThanClobbers(t *testing.T) {
 	at := h.now().Add(2 * time.Hour)
 
 	question, asked := h.askClarify(ctx, clarifyDec(router.IntentReminder,
-		router.Slots{Time: at, HasTime: true}, "напомни в 11"))
+		router.Slots{Time: at, HasTime: true}, "напомни сегодня в 11 утра"))
 	if !asked || question != "О чём напомнить?" {
 		t.Fatalf("expected the subject question, got %q asked=%v", question, asked)
 	}
@@ -573,14 +584,24 @@ func TestClarifyStepsAsideForItsOwnRequest(t *testing.T) {
 	ctx := context.Background()
 	h, _, _ := newClarifyHandler(t)
 
-	if _, asked := h.askClarify(ctx, clarifyDec(router.IntentAct, router.Slots{Text: "выключи свет в спальне"}, "выключи свет в спальне")); !asked {
-		t.Fatal("an act with no fn should be asked about")
+	// A reminder, not the act this bug was found on: since Vikunja #556 an act
+	// no longer parks anything, so it can no longer eat the turn after it.
+	if _, asked := h.askClarify(ctx, clarifyDec(router.IntentReminder, router.Slots{Text: "напомни позвонить маме"}, "напомни позвонить маме")); !asked {
+		t.Fatal("a reminder with no time should be asked about")
 	}
 	if reply, handled := h.resolveClarifyAnswer(ctx, "кто изобрёл телефон"); handled {
 		t.Fatalf("a world question must route as itself, got %q", reply)
 	}
-	if h.clarifyStore.Get(voiceDialogueID, h.now()) != nil {
-		t.Error("the parked question must be dropped, not left to eat the turn after this one")
+	// Not eating the turn is `handled == false` above, and that is the whole of
+	// #554. Since V-561 the question also SURVIVES it: a side query suspends the
+	// flow rather than ending it, so the reminder is still there and still on the
+	// attempt it was parked with.
+	q := h.clarifyStore.Get(voiceDialogueID, h.now())
+	if q == nil {
+		t.Fatal("a side query must suspend the parked question, not drop it")
+	}
+	if q.Attempts != 1 {
+		t.Errorf("a turn that was never an answer spent an attempt: %d, want 1", q.Attempts)
 	}
 }
 
@@ -613,10 +634,159 @@ func TestClarifyQuestionShapedAnswerThatFillsTheGapStillLands(t *testing.T) {
 	if _, asked := h.askClarify(ctx, clarifyDec(router.IntentReminder, router.Slots{Text: "напомни позвонить маме"}, "напомни позвонить маме")); !asked {
 		t.Fatal("expected the time question")
 	}
-	if reply, handled := h.resolveClarifyAnswer(ctx, "а что если в 11:00"); !handled || reply == clarifyGaveUp {
+	if reply, handled := h.resolveClarifyAnswer(ctx, "а что если сегодня в 11:00"); !handled || reply == clarifyGaveUp {
 		t.Fatalf("an answer that fills the gap must land, handled=%v reply=%q", handled, reply)
 	}
 	if reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour)); err != nil || len(reminders) != 1 {
 		t.Fatalf("reminder was not created: reminders=%v err=%v", reminders, err)
 	}
 }
+
+// TestUnresolvedActSaysItDoesNotKnowTheCommand — Vikunja #556. "Что сделать?"
+// has no answer he can give, so an act that matched no capability is refused in
+// one line and nothing is parked. It does not recite what she can do instead.
+func TestUnresolvedActSaysItDoesNotKnowTheCommand(t *testing.T) {
+	ctx := context.Background()
+	h, st, _ := newClarifyHandler(t)
+	// Enabled tools change nothing here: this act matched none of them.
+	if err := st.EnableTool(ctx, "uptime", []string{"true"}, false, "test", h.now()); err != nil {
+		t.Fatal(err)
+	}
+
+	reply, spoken := h.askClarify(ctx, clarifyDec(router.IntentAct, router.Slots{Text: "выключи свет"}, "выключи свет"))
+	if !spoken || reply != actNotRecognized {
+		t.Fatalf("reply = %q spoken=%v, want %q", reply, spoken, actNotRecognized)
+	}
+	if strings.Contains(reply, "uptime") {
+		t.Errorf("reply = %q, want no list of capabilities he did not ask about", reply)
+	}
+	if h.clarifyStore.Get(voiceDialogueID, h.now()) != nil {
+		t.Error("nothing to ask about, so nothing may be parked")
+	}
+}
+
+// newRoutingClarifyHandler wires the real cascade (hash embedder, no model) onto
+// the clarify handler, so a test can drive handleText end to end and see which
+// gate claimed the turn.
+func newRoutingClarifyHandler(t *testing.T) (*reactiveHandler, *store.Store) {
+	t.Helper()
+	h, st, _ := newClarifyHandler(t)
+	h.router = buildRouter(router.NewHashEmbedder(1024), h.matcher, 0.55, nil, nil)
+	h.recall = recallWiring{embedder: router.NewHashEmbedder(1024), memStore: memory.NewInMemoryStore()}
+	return h, st
+}
+
+// TestIncompleteReminderAsksInsteadOfFailing — Vikunja #557. "напомни позвонить"
+// is routed confidently and is still half a request. It used to reach applyAction,
+// fail on the missing time and park nothing, so the "в семь вечера" that followed
+// was routed as a world question and web-searched.
+func TestIncompleteReminderAsksInsteadOfFailing(t *testing.T) {
+	ctx := context.Background()
+	h, st := newRoutingClarifyHandler(t)
+
+	reply := h.handleText(ctx, "web", "напомни позвонить маме")
+	want, _ := whenQuestion(whenNoHour, 1, h.now(), "")
+	if reply != want {
+		t.Fatalf("reply = %q, want the time question %q", reply, want)
+	}
+	if h.clarifyStore.Get(dialogueIDFor(sourceText, "web"), h.now()) == nil {
+		t.Fatal("the request must be parked, or the answer has nowhere to land")
+	}
+	if reply := h.handleText(ctx, "web", "сегодня в семь вечера"); strings.Contains(reply, "нашла") {
+		t.Fatalf("the answer to her own question must not be looked up: %q", reply)
+	}
+	if reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour)); err != nil || len(reminders) != 1 {
+		t.Fatalf("the answer did not complete the reminder: reminders=%v err=%v", reminders, err)
+	}
+}
+
+// TestBareCaptureVerbAsksWhatToRecord — the other half of #557. A bare "запиши"
+// went to the resident model as chat, which agreed to a wording change nobody
+// asked for. It is a fact with no key, and that gap has a question.
+func TestBareCaptureVerbAsksWhatToRecord(t *testing.T) {
+	ctx := context.Background()
+	h, _ := newRoutingClarifyHandler(t)
+
+	reply := h.handleText(ctx, "web", "запиши")
+	want, _ := clarifyQuestionFor(dialogue.SlotKey, 1)
+	if reply != want {
+		t.Fatalf("reply = %q, want %q", reply, want)
+	}
+	if h.clarifyStore.Get(dialogueIDFor(sourceText, "web"), h.now()) == nil {
+		t.Fatal("the request must be parked so the next utterance completes it")
+	}
+}
+
+// TestACompleteTurnStillDoesNotAsk — the gate reads a missing slot, not any
+// slot, so a request she can act on must never turn into a question. Checked on
+// the decision rather than through the cascade: what is at stake is the gate's
+// condition, and driving it through the hash embedder would measure routing.
+func TestACompleteTurnStillDoesNotAsk(t *testing.T) {
+	ctx := context.Background()
+	h, _, _ := newClarifyHandler(t)
+
+	complete := []router.Decision{
+		{Intent: router.IntentReminder, Slots: router.Slots{Text: "позвонить маме", HasTime: true}, Utterance: "напомни завтра в 11 утра позвонить маме"},
+		{Intent: router.IntentFact, Slots: router.Slots{Key: "water", Value: "выпил", HasKey: true}, Utterance: "я выпил воды"},
+		{Intent: router.IntentNote, Slots: router.Slots{Text: "купить хлеб"}, Utterance: "запиши купить хлеб"},
+		{Intent: router.IntentQuery, Slots: router.Slots{Text: "что у меня сегодня"}, Utterance: "что у меня сегодня"},
+	}
+	for _, dec := range complete {
+		if gaps := missingFor(dec); len(gaps) > 0 {
+			t.Errorf("%q reads as incomplete: %v", dec.Utterance, gaps)
+		}
+		if reply, asked := h.askClarify(ctx, dec); asked {
+			t.Errorf("%q was answered with a question: %q", dec.Utterance, reply)
+		}
+	}
+}
+
+// TestTheResumedQuestionIsItsOwnSentence — V-654. The re-ask used to be spliced
+// onto the answer with a comma, so a real answer and an unrelated open question
+// read as one run-on thought and the question vanished into its tail.
+func TestTheResumedQuestionIsItsOwnSentence(t *testing.T) {
+	const resumed = "На какое время поставить напоминание?"
+	cases := []struct {
+		name  string
+		reply string
+		want  string
+	}{
+		{
+			// The measured line, shortened. Two sentences, and the question keeps
+			// its capital.
+			name:  "a statement keeps its full stop",
+			reply: "Вайфай пароль лежит в ящике стола.",
+			want:  "Вайфай пароль лежит в ящике стола. " + resumed,
+		},
+		{
+			name:  "a reply with no terminator is given one",
+			reply: "Вайфай пароль лежит в ящике стола",
+			want:  "Вайфай пароль лежит в ящике стола. " + resumed,
+		},
+		{
+			// She sometimes answers a side query by asking him to say it again.
+			// Two questions, and neither may swallow the other.
+			name:  "a question keeps its mark",
+			reply: "Можешь переформулировать?",
+			want:  "Можешь переформулировать? " + resumed,
+		},
+		{
+			name:  "an ellipsis is already an ending",
+			reply: "Не уверена…",
+			want:  "Не уверена… " + resumed,
+		},
+		{
+			name:  "a resume with no answer in front of it is just the question",
+			reply: "",
+			want:  resumed,
+		},
+	}
+	for _, tc := range cases {
+		if got := withResumed(tc.reply, resumed); got != tc.want {
+			t.Errorf("%s: withResumed(%q) = %q, want %q", tc.name, tc.reply, got, tc.want)
+		}
+	}
+	if got := withResumed("Готово.", ""); got != "Готово." {
+		t.Errorf("nothing to resume must leave the reply alone, got %q", got)
+	}
+}

cmd/mavend/clarifytemplates.go

@@ -40,6 +40,9 @@ var clarifyQuestionVariants = map[dialogue.Slot][]string{
 		"Что именно отметить?",
 		"Назови, что записать — например, «выпил воды».",
 	},
+	// Not spoken since Vikunja #556: askClarify answers actNotRecognized for a
+	// missing capability rather than asking. Kept because clarifyQuestion still
+	// reports the gap, and a re-ask deck with a hole in it is harder to read.
 	dialogue.SlotFn: {
 		"Что сделать?",
 		"Какое действие выполнить?",
@@ -47,6 +50,49 @@ var clarifyQuestionVariants = map[dialogue.Slot][]string{
 	},
 }
 
+// clarifyResumedVariants — the wording for a question coming BACK after a side
+// query took the turn away from it (Vikunja #561).
+//
+// It is not the first question again. "Когда?" works in the same breath as
+// "напомни позвонить маме", because the thing it is about was just said. After
+// a turn about the weather in Rome it does not: he has been thinking about
+// something else, and a bare "Когда?" asks him to remember what she is holding.
+// So the resumed form names the request — "напоминание", "заметка" — and the
+// first form stays short.
+//
+// One wording per slot, not a rotation and not an attempt ladder. A resume does
+// not spend an attempt (that is the point of suspending rather than re-asking),
+// so there is no attempt number to vary on, and this line is heard once per
+// interruption rather than repeatedly.
+//
+// Persona holds: infinitive, so no gender agreement, "ты" nowhere needed, no pet
+// names.
+var clarifyResumedVariants = map[dialogue.Slot]string{
+	dialogue.SlotTime: "На какое время поставить напоминание?",
+	dialogue.SlotText: "Так о чём напомнить?",
+	dialogue.SlotKey:  "Так что записать?",
+	dialogue.SlotFn:   "Так какое действие выполнить?",
+}
+
+// clarifyResumedFor gives the resumed wording for a slot. ("", false) when the
+// slot has none, and the caller then resumes nothing rather than inventing a
+// question — a flow it cannot re-ask is one it must not claim to be holding.
+func clarifyResumedFor(slot dialogue.Slot) (string, bool) {
+	q, ok := clarifyResumedVariants[slot]
+	return q, ok
+}
+
+// actNotRecognized is what an act she cannot run gets (Vikunja #556).
+//
+// The deck used to ask "Что сделать?" instead. That question has no answer he
+// can give: he already said what he wanted, and nothing he repeats will match a
+// capability that is not there. So she asked, failed, asked again and gave up —
+// three turns spent on one refusal. She says it once now, and parks nothing.
+//
+// It does not recite the allowlist. A list of names he did not ask about is not
+// an answer to the thing he did ask about.
+const actNotRecognized = "Такую команду я не знаю."
+
 // clarifyQuestionFor picks the wording for this attempt. attempt is 1-based, as
 // PendingQuestion.Attempts counts it; anything past the list uses the last and
 // most explicit phrasing rather than wrapping round to the short one, because

cmd/mavend/confirm.go

@@ -3,9 +3,13 @@ package main
 import (
 	"context"
 	"log"
+	"slices"
+	"sort"
 	"strings"
 	"time"
+	"unicode"
 
+	"github.com/kami/maven/internal/lexicon"
 	"github.com/kami/maven/internal/phraser"
 	"github.com/kami/maven/internal/router"
 )
@@ -20,6 +24,14 @@ type pendingHexisExec struct {
 	entityID     string
 	displayName  string
 	expiry       time.Time
+
+	// correlationID — the id the proposing turn minted for this action. A
+	// confirm arrives on a later turn with a context of its own, so without
+	// carrying it here the execution recorded a fresh id and no causation at
+	// all, and the resolve, the discovery and the thing they authorised sat in
+	// the trace as unrelated calls. The contract mints one id per action, and
+	// the action began when she asked.
+	correlationID string
 }
 
 // pendingRoutineConfirm — a proposed routine awaiting a spoken y/n to become
@@ -57,10 +69,18 @@ func (h *reactiveHandler) park(fn string, args []string, phrase string) {
 // resolveConfirm interprets an utterance as the answer to a parked destructive
 // act OR a parked routine proposal. Returns (reply, true) when it consumed the
 // utterance as a y/n answer; ("", false) when there's nothing pending (or the
-// parked act expired), so the caller routes the utterance normally. An
-// unrecognised answer cancels the pending and routes normally — a confirm that
-// can't be answered clearly is safer abandoned than left armed.
+// parked act expired), so the caller routes the utterance normally.
+//
+// An utterance that is not clearly yes or no is not an answer at all, so it is
+// handed straight back and the pending stays parked until it expires (V-567).
+// This resolver runs before routing and holds the most dangerous trigger on the
+// box; it may only claim a turn it is certain about.
 func (h *reactiveHandler) resolveConfirm(ctx context.Context, text string) (string, bool) {
+	verdict := classifyConfirm(text)
+	if verdict == confirmUnknown {
+		return "", false
+	}
+
 	h.mu.Lock()
 	defer h.mu.Unlock()
 
@@ -68,16 +88,12 @@ func (h *reactiveHandler) resolveConfirm(ctx context.Context, text string) (stri
 		if !r.claim() {
 			continue
 		}
-		// The slot is already cleared by claim(): every branch below drops the
-		// pending, including the unclear one — a confirm that can't be
-		// answered clearly is safer abandoned than left armed.
-		switch classifyConfirm(text) {
+		// The slot is already cleared by claim().
+		switch verdict {
 		case confirmYes:
 			return r.yes(), true
-		case confirmNo:
-			return r.no(), true
 		default:
-			return "", false
+			return r.no(), true
 		}
 	}
 	return "", false
@@ -136,7 +152,11 @@ func (h *reactiveHandler) confirmResolvers(ctx context.Context) []confirmResolve
 				return hx != nil && !h.now().After(hx.expiry)
 			},
 			yes: func() string {
-				return h.execHexis(ctx, hx.capabilityID, hx.capName, hx.entityID, hx.displayName)
+				execCtx := ctx
+				if hx.correlationID != "" {
+					execCtx = withCorrelationID(execCtx, hx.correlationID)
+				}
+				return h.execHexis(execCtx, hx.capabilityID, hx.capName, hx.entityID, hx.displayName)
 			},
 			no: func() string { return phraser.C(phraser.ConfirmCancelled, nil) },
 		},
@@ -194,23 +214,92 @@ const (
 	confirmNo
 )
 
-// classifyConfirm reads a short ru/en yes-or-no answer. Substring match on the
-// stems so inflections/fillers ("да, давай", "нет, отмени") still land.
+// confirmWords are the two closed sets, tokenized once and ordered
+// longest-first so "не надо" is read before "нет" could claim any of it.
+var (
+	confirmYesPhrases = confirmPhrases(lexicon.ConfirmYes())
+	confirmNoPhrases  = confirmPhrases(lexicon.ConfirmNo())
+)
+
+// confirmPhrases splits each lexicon member into tokens and sorts the result
+// longest-first, so a walk that tries them in order matches the longest member
+// that fits.
+func confirmPhrases(words []string) [][]string {
+	out := make([][]string, 0, len(words))
+	for _, w := range words {
+		if toks := confirmTokens(w); len(toks) > 0 {
+			out = append(out, toks)
+		}
+	}
+	sort.SliceStable(out, func(i, j int) bool { return len(out[i]) > len(out[j]) })
+	return out
+}
+
+// confirmTokens splits an utterance into lowercase word tokens. Punctuation and
+// spacing are separators; an apostrophe is not, because "don't" is one word.
+func confirmTokens(text string) []string {
+	return strings.FieldsFunc(strings.ToLower(text), func(r rune) bool {
+		if r == '\'' || r == '’' {
+			return false
+		}
+		return !unicode.IsLetter(r) && !unicode.IsDigit(r)
+	})
+}
+
+// classifyConfirm reads a short ru/en yes-or-no answer to a parked confirm.
+//
+// The whole utterance must consist of confirmation words and filler, matched as
+// whole tokens against the closed lexicon sets. Anything else is
+// confirmUnknown, which leaves the confirm parked and routes the turn — see
+// resolveConfirm. Both halves of that are the fix for V-567: this used to be a
+// substring test over bare stems, so "погода", "дальше", "надо" and "давление"
+// all read as "да", and "покажи" and "около" read as "ок". A parked destructive
+// act fired on a question about the weather.
+//
+// Requiring the WHOLE utterance is the second half. A leading confirm word does
+// not make a sentence an answer: "давай посмотрим погоду" opens a request, and
+// the only safe reading of a sentence that carries its own subject is that he
+// moved on. Guessing wrong here executes something; guessing wrong the other way
+// asks again.
 func classifyConfirm(text string) confirmVerdict {
-	t := strings.ToLower(strings.TrimSpace(text))
-	// negatives first — "не надо" contains no "да", but check no-stems before
-	// yes so a leading "нет" isn't shadowed.
-	for _, no := range []string{"нет", "не надо", "отмен", "стоп", "no", "cancel", "stop", "don't"} {
-		if strings.Contains(t, no) {
+	tokens := confirmTokens(text)
+	if len(tokens) == 0 {
+		return confirmUnknown
+	}
+	verdict := confirmUnknown
+	for i := 0; i < len(tokens); {
+		// Negatives first: "не надо" and "не хочу" open with a token that is
+		// not itself an answer, and a yes hit must never shadow them.
+		if n := matchConfirm(confirmNoPhrases, tokens[i:]); n > 0 {
 			return confirmNo
 		}
+		if n := matchConfirm(confirmYesPhrases, tokens[i:]); n > 0 {
+			verdict, i = confirmYes, i+n
+			continue
+		}
+		if lexicon.IsFillerParticle(tokens[i]) {
+			i++
+			continue
+		}
+		// A word that is neither an answer nor filler carries a subject of its
+		// own, so this utterance is not an answer to her question.
+		return confirmUnknown
 	}
-	for _, yes := range []string{"да", "ага", "давай", "подтвер", "конечно", "yes", "yeah", "yep", "confirm", "ок", "okay", "ok"} {
-		if strings.Contains(t, yes) {
-			return confirmYes
+	return verdict
+}
+
+// matchConfirm reports the length of the longest phrase matching at the head of
+// tokens, or 0.
+func matchConfirm(phrases [][]string, tokens []string) int {
+	for _, p := range phrases {
+		if len(p) > len(tokens) {
+			continue
+		}
+		if slices.Equal(p, tokens[:len(p)]) {
+			return len(p)
 		}
 	}
-	return confirmUnknown
+	return 0
 }
 
 // actPhrase renders "fn arg1 arg2" for the confirm prompt.

cmd/mavend/confirm_test.go

@@ -0,0 +1,103 @@
+package main
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/tool"
+)
+
+// TestClassifyConfirmRejectsSubstrings — V-567. The old matcher tested bare
+// stems with strings.Contains, so every word below answered a question she had
+// asked about something else: "погода", "дальше", "надо" and "давление" carry
+// "да"; "покажи", "около" and "окно" carry "ок". A parked destructive act fired
+// on a question about the weather.
+func TestClassifyConfirmRejectsSubstrings(t *testing.T) {
+	for _, text := range []string{
+		"погода",
+		"какая погода",
+		"что дальше",
+		"надо ещё",
+		"покажи заметки",
+		"около окна",
+		"давление",
+		"давай посмотрим погоду",
+		"не забудь купить хлеб",
+		"окно открыто",
+		"стоит ли брать зонт",
+		"",
+	} {
+		if got := classifyConfirm(text); got != confirmUnknown {
+			t.Errorf("classifyConfirm(%q) = %v, want confirmUnknown", text, got)
+		}
+	}
+}
+
+// TestClassifyConfirmAcceptsAnswers keeps every genuine answer the substring
+// matcher accepted, and pins the pair the fix could most easily get wrong:
+// "надо" is not an answer and "не надо" is the opposite of one.
+func TestClassifyConfirmAcceptsAnswers(t *testing.T) {
+	yes := []string{"да", "Да!", "ага", "давай", "да, давай", "конечно", "подтверждаю", "ну да", "yes", "yeah", "ok", "okay", "confirm"}
+	no := []string{"нет", "Нет.", "не надо", "не нужно", "не сейчас", "отмена", "отмени", "стоп", "нет, отмени", "no", "nope", "cancel", "stop", "don't"}
+
+	for _, text := range yes {
+		if got := classifyConfirm(text); got != confirmYes {
+			t.Errorf("classifyConfirm(%q) = %v, want confirmYes", text, got)
+		}
+	}
+	for _, text := range no {
+		if got := classifyConfirm(text); got != confirmNo {
+			t.Errorf("classifyConfirm(%q) = %v, want confirmNo", text, got)
+		}
+	}
+}
+
+// TestUnrelatedTurnLeavesConfirmParked — the whole point of V-567. An utterance
+// that is not an answer must not execute the parked act, must not consume the
+// turn, and must not disarm the confirm either: the answer he has not given yet
+// is still answerable until it expires.
+func TestUnrelatedTurnLeavesConfirmParked(t *testing.T) {
+	ctx := context.Background()
+	st := newTestStore(t)
+	api := ipc.NewStoreAPI(st)
+	now := time.Date(2026, 8, 6, 9, 0, 0, 0, time.UTC)
+	h := &reactiveHandler{
+		api:       api,
+		dataStore: st,
+		now:       func() time.Time { return now },
+		tools:     tool.NewExecutor(api, time.Second),
+	}
+
+	marker := filepath.Join(t.TempDir(), "destructive-tool-ran")
+	if err := st.EnableTool(ctx, "delete_backups", []string{"touch", marker}, true, "test", h.now()); err != nil {
+		t.Fatal(err)
+	}
+	h.park("delete_backups", nil, "delete_backups")
+
+	if reply, handled := h.resolveConfirm(ctx, "какая погода"); handled {
+		t.Fatalf("the weather question was consumed as a confirm: %q", reply)
+	}
+	if _, err := os.Stat(marker); !os.IsNotExist(err) {
+		t.Fatalf("the parked destructive command ran on an unrelated turn: %v", err)
+	}
+	if h.pending == nil {
+		t.Fatal("the confirm was disarmed by a turn that did not answer it")
+	}
+
+	// It is still answerable, and answering it still runs the act.
+	reply, handled := h.resolveConfirm(ctx, "да")
+	if !handled || !strings.Contains(reply, "готово") {
+		t.Fatalf("the still-parked confirm did not resolve: handled=%v reply=%q", handled, reply)
+	}
+	if _, err := os.Stat(marker); err != nil {
+		t.Fatalf("confirmed destructive command did not run: %v", err)
+	}
+	if h.pending != nil {
+		t.Fatal("the confirm stayed parked after being answered")
+	}
+}

cmd/mavend/crawls.go

@@ -173,6 +173,13 @@ func (a *crawlFetcher) Get(ctx context.Context, u string) (*crawl.Response, erro
 		case errors.Is(err, webfetch.ErrBlocked), errors.Is(err, webfetch.ErrPrivate), errors.Is(err, webfetch.ErrScheme):
 			return nil, fmt.Errorf("%w: %v", crawl.ErrFetchRefused, err)
 		case errors.Is(err, webfetch.ErrStatus):
+			// Carry the code across the seam. The crawler needs to tell a 5xx
+			// from a 404 to decide what a failed robots.txt means, and it must
+			// not learn that by reading this sentence.
+			var se *webfetch.StatusError
+			if errors.As(err, &se) {
+				return nil, &crawl.StatusError{Code: se.Code}
+			}
 			return nil, fmt.Errorf("%w: %v", crawl.ErrFetchStatus, err)
 		}
 		return nil, err

cmd/mavend/dayplan_test.go

@@ -4,12 +4,14 @@ import (
 	"context"
 	"database/sql"
 	"errors"
+	"fmt"
 	"strings"
 	"testing"
 	"time"
 
 	"github.com/kami/maven/internal/calendar"
 	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/morning"
 	"github.com/kami/maven/internal/phraser"
 	"github.com/kami/maven/internal/router"
 	"github.com/kami/maven/internal/store"
@@ -111,6 +113,88 @@ func TestQueryDayPlanRestOfDayWhenNothingIsLeft(t *testing.T) {
 	}
 }
 
+// "что дальше?" rebuilds the plan off the wire and renders it here, and the
+// instants on it carry the zone the core read them in — a calendar fact's Ts
+// and a reminder's FireTs are UTC out of the store. Read raw, the recital named
+// the store's clock instead of his (V-614). The asking clock is three hours off
+// whatever this machine runs in, so the assertion holds under TZ=UTC too.
+func TestQueryDayPlanRestOfDayReadsHisClock(t *testing.T) {
+	_, off := time.Now().Zone()
+	away := time.FixedZone("away", off+3*60*60)
+	stored := time.Date(2026, 8, 3, 8, 0, 0, 0, time.UTC)
+
+	h := &reactiveHandler{
+		api: &planAPI{plan: ipc.DayPlan{
+			Date:  time.Date(2026, 8, 3, 0, 0, 0, 0, time.UTC),
+			Items: []ipc.DayPlanItem{{At: stored, Text: "позвонить маме", Kind: "reminder"}},
+		}},
+		now: func() time.Time { return time.Date(2026, 8, 3, 9, 0, 0, 0, away) },
+	}
+	reply, ok := h.queryDayPlan(context.Background(), &queryTurn{
+		dec: router.Decision{Intent: router.IntentQuery, Utterance: "что дальше?"},
+	})
+	if !ok {
+		t.Fatal("expected the plan source to claim it")
+	}
+	if want := stored.In(away).Format("15:04"); !strings.Contains(reply, want) {
+		t.Errorf("the reminder is not read in his clock (%s): %q", want, reply)
+	}
+	if bad := stored.Format("15:04"); strings.Contains(reply, bad) {
+		t.Errorf("the reminder is read in the store's zone (%s): %q", bad, reply)
+	}
+}
+
+// The defect V-618 fixes, at the handler: asked at 04:45 the trim removes
+// nothing, because the whole day is still ahead. She read 43 entries aloud as
+// one sentence. The zone is three hours off UTC so the test also fails under
+// TZ=UTC if the rendering ever slips zones.
+func TestQueryDayPlanCapsWhatItReadsAloud(t *testing.T) {
+	zone := time.FixedZone("MSK", 3*60*60)
+	mid := time.Date(2026, 8, 3, 0, 0, 0, 0, zone)
+	plan := ipc.DayPlan{Date: mid, Spoken: "план на 03.08.2026: …"}
+	for i := 0; i < 43; i++ {
+		plan.Items = append(plan.Items, ipc.DayPlanItem{
+			At:   mid.Add(time.Duration(345+i*20) * time.Minute), // 05:45 onward
+			Text: fmt.Sprintf("пункт %d", i),
+			Kind: "event",
+		})
+	}
+	h := &reactiveHandler{api: &planAPI{plan: plan}, now: func() time.Time {
+		return time.Date(2026, 8, 3, 4, 45, 0, 0, zone)
+	}}
+	reply, ok := h.queryDayPlan(context.Background(), &queryTurn{
+		dec: router.Decision{Intent: router.IntentQuery, Utterance: "что дальше?"},
+	})
+	if !ok {
+		t.Fatal("expected the plan source to claim it")
+	}
+	if n := strings.Count(reply, "пункт "); n != morning.NextSpoken {
+		t.Errorf("read %d entries aloud, want %d: %q", n, morning.NextSpoken, reply)
+	}
+	if !strings.HasPrefix(reply, "дальше: 05:45 — пункт 0;") {
+		t.Errorf("the next thing is not first: %q", reply)
+	}
+	// The rest is counted, not silently dropped.
+	if !strings.Contains(reply, "и ещё 40 дел до конца дня.") {
+		t.Errorf("the sentence hides that the day goes on: %q", reply)
+	}
+}
+
+// "что у меня сегодня?" is the whole day and is not narrowed. It carries no
+// plan word, so the plan source declines it and the calendar listing answers —
+// asserted here beside the cap so the two questions cannot drift together.
+func TestWholeDayQuestionIsNotTheRestOfTheDay(t *testing.T) {
+	if router.IsDayPlanQuery("что у меня сегодня?") {
+		t.Error("the plan source claims the whole-day question")
+	}
+	if !router.IsDayPlanQuery("что дальше?") {
+		t.Error("the plan source stopped claiming the rest-of-day question")
+	}
+	if router.IsRestOfDayQuery("какие планы на сегодня?") {
+		t.Error("the whole-day plan question got narrowed to the rest of the day")
+	}
+}
+
 // A question that is not about the plan must fall through, or the plan buries
 // the calendar listing and the weather behind it.
 func TestQueryDayPlanPassesOnEverythingElse(t *testing.T) {

cmd/mavend/decisiontrace.go

@@ -0,0 +1,133 @@
+// mavend/decisiontrace.go — the daemon's half of the per-turn decision record.
+//
+// V-564. The router says what the cascade did (internal/router/decisiontrace.go);
+// this file covers the two claimant sets that live in the daemon: the stateful
+// resolvers that run BEFORE routing and pre-empt it unconditionally, and the
+// query source chain that runs after. Those two are where the arbitration is
+// least visible, because both are a hardcoded order of functions that each
+// answer "is this mine?" alone and none of which answers "is this more mine
+// than yours?" (V-558).
+//
+// Recording changes no route. Every helper here is a no-op on a context with no
+// record, which is what every test that does not ask for one gets.
+package main
+
+import (
+	"context"
+
+	"github.com/kami/maven/internal/decision"
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/router"
+)
+
+// preRouteLadder — the resolvers runTurn offers the utterance to before the
+// router sees it, in the order they get their say. Kept here as a roster rather
+// than derived from the code, so a resolver that returns early and skips the
+// rest still leaves the rest NAMED in the record: a claimant that never looked
+// and one that looked and passed are the distinction the ordering hides, and
+// they are the difference between a bug in the ladder and a bug in a resolver.
+//
+// Adding a step to runTurn means adding its name here. Nothing enforces that,
+// and nothing should: a missing name costs one line of the record, while a
+// check that walks the ladder would have to run the ladder.
+var preRouteLadder = []string{
+	"confirm", "clarify-answer", "quiet-toggle", "snooze", "ack", "repair",
+	"repair-negative", "ordinal",
+}
+
+// notePreRoute records one rung of that ladder and passes its verdict through
+// unchanged, so the call site stays the single `if handled` it already was.
+func notePreRoute(ctx context.Context, name string, handled bool) bool {
+	rec := decision.From(ctx)
+	if rec == nil {
+		return handled
+	}
+	if handled {
+		rec.Note(decision.Claim{
+			Stage: decision.StagePreRoute, Claimant: name, Outcome: decision.Won,
+			Reason: "it pre-empted routing, so the router never saw this turn",
+		})
+		return handled
+	}
+	rec.Note(decision.Claim{
+		Stage: decision.StagePreRoute, Claimant: name, Outcome: decision.Declined,
+		Reason: "nothing of its own was pending",
+	})
+	return handled
+}
+
+// noteTerminal records whoever actually produced the reply, but only if the
+// turn is still unclaimed. A route decides the intent; it does not answer, and
+// on a thinned route or a plain act nothing downstream keeps a scoreboard. So
+// the record would otherwise close with an empty winner, which reads as a lost
+// turn instead of an asked question.
+func noteTerminal(ctx context.Context, claimant string, intent router.Intent, reason string) {
+	decision.From(ctx).NoteIfUnclaimed(decision.Claim{
+		Stage: decision.StageAction, Claimant: claimant,
+		Intent: string(intent), Reason: reason,
+	})
+}
+
+// noteMerge records the follow-up merge, which is the one claimant that edits
+// the winning decision instead of taking the turn from it. It is compared on
+// the four slots the merge can fill, because a Decision holds a slice and is
+// not comparable.
+func noteMerge(ctx context.Context, before, after router.Decision) {
+	rec := decision.From(ctx)
+	if rec == nil {
+		return
+	}
+	changed := before.Slots.HasTime != after.Slots.HasTime ||
+		before.Slots.HasKey != after.Slots.HasKey ||
+		before.Slots.HasFn != after.Slots.HasFn ||
+		before.Slots.Text != after.Slots.Text ||
+		before.Intent != after.Intent
+	if !changed {
+		rec.Note(decision.Claim{
+			Stage: decision.StageMerge, Claimant: "follow-up-merge", Outcome: decision.Declined,
+			Reason: "no slot of this turn was left for a previous one to fill",
+		})
+		return
+	}
+	rec.Note(decision.Claim{
+		Stage: decision.StageMerge, Claimant: "follow-up-merge", Intent: string(after.Intent),
+		Outcome: decision.Merged, Reason: "filled this turn's gaps from the previous turn",
+	})
+}
+
+// turnDecisionsFn — the reader mavweb gets, or nil when voice was never wired.
+// Same shape as intakeEventsFn: the daemon holds the ring, the IPC layer only
+// converts it.
+func turnDecisionsFn(w *voiceWiring) func(int) []ipc.TurnDecision {
+	if w == nil || w.handler == nil || w.handler.decisions == nil {
+		return nil
+	}
+	ring := w.handler.decisions
+	return func(n int) []ipc.TurnDecision {
+		recs := ring.Recent(n)
+		out := make([]ipc.TurnDecision, 0, len(recs))
+		for _, rec := range recs {
+			claims := make([]ipc.TurnClaim, 0, len(rec.Claims))
+			for _, c := range rec.Claims {
+				claims = append(claims, ipc.TurnClaim{
+					Stage: c.Stage, Claimant: c.Claimant, Intent: c.Intent,
+					Score: c.Score, HasScore: c.HasScore,
+					Outcome: c.Outcome, Reason: c.Reason,
+				})
+			}
+			out = append(out, ipc.TurnDecision{
+				Ts: rec.Ts, Utterance: rec.Utterance, Winner: rec.Winner, Claims: claims,
+			})
+		}
+		return out
+	}
+}
+
+// querySourceNames — the query chain's roster, in chain order.
+func querySourceNames() []string {
+	names := make([]string, len(querySources))
+	for i, src := range querySources {
+		names[i] = src.name
+	}
+	return names
+}

cmd/mavend/decisiontrace_test.go

@@ -0,0 +1,163 @@
+package main
+
+import (
+	"context"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/decision"
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/memory"
+	"github.com/kami/maven/internal/router"
+	"github.com/kami/maven/internal/tool"
+	"github.com/kami/maven/internal/voice"
+)
+
+// traceHandler — a handler with the decision ring wired, the same shape the
+// daemon builds in wireVoice.
+func traceHandler(t *testing.T, ring *decision.Ring) *reactiveHandler {
+	t.Helper()
+	st := newTestStore(t)
+	api := ipc.NewStoreAPI(st)
+	now := time.Now()
+	emb := router.NewHashEmbedder(1024)
+	return &reactiveHandler{
+		api:       api,
+		recall:    recallWiring{embedder: emb, memStore: memory.NewInMemoryStore()},
+		router:    buildRouter(emb, tool.NewMatcher(api), 0.55, nil, nil),
+		replier:   voice.NewStubReplier(),
+		now:       func() time.Time { return now },
+		dataStore: st,
+		decisions: ring,
+	}
+}
+
+// findClaim — the first claim for a claimant, or nil.
+func findClaim(rec *decision.Record, claimant string) *decision.Claim {
+	for i := range rec.Claims {
+		if rec.Claims[i].Claimant == claimant {
+			return &rec.Claims[i]
+		}
+	}
+	return nil
+}
+
+// TestTurnRecordNamesWinnerAndLosers — the point of V-564. A turn a stage-0
+// grammar claims must leave a record naming that grammar as the winner, naming
+// a pre-route resolver that declined, and naming the routing engines that were
+// never reached at all. The last of those is the fact the hardcoded ordering
+// hides: "classifier" absent from the record and "classifier" never asked read
+// the same to a human, and only one of them is the truth.
+func TestTurnRecordNamesWinnerAndLosers(t *testing.T) {
+	ring := decision.NewRing()
+	h := traceHandler(t, ring)
+
+	reply := h.handleText(context.Background(), "web", "сколько сейчас времени")
+	if reply == "" {
+		t.Fatal("turn produced no reply")
+	}
+
+	recs := ring.Recent(5)
+	if len(recs) != 1 {
+		t.Fatalf("want 1 record, got %d", len(recs))
+	}
+	rec := recs[0]
+	if rec.Utterance != "сколько сейчас времени" {
+		t.Errorf("utterance = %q", rec.Utterance)
+	}
+	if !strings.HasPrefix(rec.Winner, "stage0:") {
+		t.Errorf("want a stage-0 grammar as the winner, got %q", rec.Winner)
+	}
+
+	// A loser that examined the turn: the confirm resolver ran first and had
+	// nothing pending.
+	confirm := findClaim(rec, "confirm")
+	if confirm == nil || confirm.Outcome != decision.Declined {
+		t.Errorf("confirm claim = %+v, want a decline", confirm)
+	}
+
+	// A loser that never looked: stage 0 answered, so neither routing engine
+	// was reached.
+	for _, name := range []string{"llm-router", "classifier"} {
+		c := findClaim(rec, name)
+		if c != nil && c.Outcome == decision.Won {
+			t.Errorf("%s cannot have won a stage-0 turn: %+v", name, c)
+		}
+	}
+
+	// And every rung of the ladder below the winner is named, not omitted.
+	for _, name := range preRouteLadder {
+		if findClaim(rec, name) == nil {
+			t.Errorf("ladder rung %q is missing from the record", name)
+		}
+	}
+}
+
+// TestRecordingDoesNotChangeTheReply — instrumentation, so a turn with the ring
+// wired and the same turn without it must answer identically. If this ever
+// fails, a claim site is doing more than noting.
+func TestRecordingDoesNotChangeTheReply(t *testing.T) {
+	for _, utt := range []string{
+		"сколько сейчас времени",
+		"запиши что я пил воду",
+		"что у меня сегодня",
+	} {
+		withRing := traceHandler(t, decision.NewRing()).handleText(context.Background(), "web", utt)
+		without := traceHandler(t, nil).handleText(context.Background(), "web", utt)
+		if withRing != without {
+			t.Errorf("%q: recorded reply %q != unrecorded %q", utt, withRing, without)
+		}
+	}
+}
+
+// TestQueryChainRecordsWhoWasNeverAsked — a query source below the claimant is
+// never consulted, and the record must say so rather than leave it out. This is
+// the arm that would have explained the Rome misroute in one read.
+func TestQueryChainRecordsWhoWasNeverAsked(t *testing.T) {
+	ring := decision.NewRing()
+	h := traceHandler(t, ring)
+	h.handleText(context.Background(), "web", "что у меня сегодня")
+
+	rec := ring.Recent(1)[0]
+	var asked, never int
+	for _, c := range rec.Claims {
+		if c.Stage != decision.StageQuery {
+			continue
+		}
+		if c.Outcome == decision.NeverAsked {
+			never++
+		} else {
+			asked++
+		}
+	}
+	if asked == 0 {
+		t.Fatal("no query source reported at all")
+	}
+	if never == 0 {
+		t.Fatal("no query source was recorded as never asked; the chain cannot have run to the end")
+	}
+	if got := len(querySourceNames()); asked+never != got {
+		t.Errorf("record covers %d of %d query sources", asked+never, got)
+	}
+}
+
+// TestTurnDecisionsFnConvertsTheRing — the IPC read path. Nil when voice was
+// never wired, because a box with no turns is an empty page and not an error.
+func TestTurnDecisionsFnConvertsTheRing(t *testing.T) {
+	if fn := turnDecisionsFn(nil); fn != nil {
+		t.Error("no wiring should mean no reader")
+	}
+	ring := decision.NewRing()
+	h := traceHandler(t, ring)
+	h.handleText(context.Background(), "web", "сколько сейчас времени")
+
+	fn := turnDecisionsFn(&voiceWiring{handler: h})
+	if fn == nil {
+		t.Fatal("wired handler produced no reader")
+	}
+	out := fn(10)
+	if len(out) != 1 || out[0].Winner == "" || len(out[0].Claims) == 0 {
+		t.Fatalf("conversion lost the record: %+v", out)
+	}
+}

cmd/mavend/dialogue_contract_test.go

@@ -0,0 +1,807 @@
+package main
+
+import (
+	"bytes"
+	"context"
+	"log"
+	"os"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/dialogue"
+	"github.com/kami/maven/internal/memory"
+	"github.com/kami/maven/internal/router"
+	"github.com/kami/maven/internal/store"
+)
+
+// Dialogue contract tests (V-563, child of V-558).
+//
+// Every other clarify test is single-shot: one ask, one answer, one assertion.
+// Three bugs of the same family shipped in two days that way — V-554 (a parked
+// question ate the three turns after it), V-557 (a confidently routed but
+// incomplete reminder parked nothing, so the answer was web-searched) and the
+// Rome case in V-558 (a side question was eaten as the time answer). None of
+// them is visible in one turn. The dialogue path is a state machine, so it can
+// be enumerated instead: whole traces, each with a per-turn expectation and an
+// expected END state — what was written to the store, and what is still parked.
+//
+// Two rules for the rows below.
+//
+// Where today's behaviour is correct, it is asserted. Where it is WRONG, the row
+// carries the CORRECT expectation and is skipped with the Vikunja id that will
+// unskip it. A weakened expectation would be worse than no row: it would pin the
+// bug as the contract.
+//
+// Everything runs on the offline floor — hash embedder, no llama-server, no
+// ONNX, StubDateTimeParser. That has one consequence worth knowing before
+// reading a fire time here: the stub reads "в 11:00" and "через час" and does
+// not read "на 9" or "на завтра", so a trace that needs those is noted where it
+// sits.
+
+// claim — which claimant consumed an utterance. Not asserted: it is derived from
+// the log lines the daemon already emits and printed on every failure, because
+// "the reply differed" does not distinguish a wrong claimant from wrong copy,
+// and that distinction is the whole point of V-558.
+type claim struct {
+	utterance string
+	steps     []string
+}
+
+func (c claim) String() string { return c.utterance + " ⇒ " + strings.Join(c.steps, " → ") }
+
+// claimMarkers — log fragment to claimant name, in the order runTurn checks
+// them. The fragments are the daemon's own words (clarify.go, repair.go,
+// voice.go); a rename there shows up here as an "unclaimed" step rather than a
+// silent mislabel.
+var claimMarkers = []struct{ fragment, name string }{
+	{"parked question expired", "clarify:expired"},
+	{"is its own request", "clarify:stepped-aside"},
+	{"gave up on", "clarify:gave-up"},
+	{"did not fill", "clarify:re-ask"},
+	{"one gap filled", "clarify:ask-second-gap"},
+	{"asked about", "clarify:ask"},
+	{"repair —", "repair"},
+	{"route result: intent=", "route"},
+}
+
+// claimsOf reads the turn's log output and names the claimants that touched it.
+func claimsOf(utterance, logged string) claim {
+	c := claim{utterance: utterance}
+	for _, line := range strings.Split(logged, "\n") {
+		for _, m := range claimMarkers {
+			if strings.Contains(line, m.fragment) {
+				name := m.name
+				if m.name == "route" {
+					name = "route:" + intentInLine(line)
+				}
+				c.steps = append(c.steps, name)
+				break
+			}
+		}
+	}
+	if len(c.steps) == 0 {
+		c.steps = []string{"unclaimed"}
+	}
+	return c
+}
+
+func intentInLine(line string) string {
+	_, rest, ok := strings.Cut(line, "intent=")
+	if !ok {
+		return "?"
+	}
+	intent, _, _ := strings.Cut(rest, " ")
+	return intent
+}
+
+// parkedWant — the question that must be armed after a turn. Attempt matters:
+// a claimant that spends a retry on an utterance that was never an answer is
+// exactly the V-554 shape, and the count is the only place it shows.
+type parkedWant struct {
+	slot    dialogue.Slot
+	attempt int
+	// carries — a substring the parked utterance must still hold, so a re-park
+	// that lost the answered subject fails here rather than three turns later.
+	carries string
+}
+
+// turn — one utterance and everything that must be true right after it.
+type turn struct {
+	say string
+	// wait — the clock moves this far BEFORE the utterance. The only way to
+	// reach the TTL without sleeping.
+	wait time.Duration
+	// question — the reply must be exactly this clarify question, worded for
+	// this attempt. Zero slot ⇒ not checked.
+	question dialogue.Slot
+	attempt  int
+	// gap — which part of the time she is asking about, for a SlotTime question
+	// (V-579). Zero value is the missing hour, which is what she asks first.
+	gap whenGap
+	// took — the words of the PREVIOUS turn that this ask must acknowledge
+	// before asking again (V-593). Empty ⇒ the ask carries no acknowledgement,
+	// which is right for a first ask and for an answer that moved nothing.
+	took string
+	// differs — this reply must not be byte-identical to the one before it. Set
+	// on a re-ask whose turn moved the request forward (V-593).
+	differs    bool
+	contains   []string
+	notContain []string
+	// noQuestion — the reply must not be any clarify question. Used where the
+	// correct behaviour is known but her wording for it is not written yet: a
+	// cancel must not be answered with another question, whatever it does say.
+	noQuestion bool
+	expired    bool // the reply must open with the TTL notice
+	// parked — what is armed after the turn. nil ⇒ nothing may be armed.
+	parked *parkedWant
+}
+
+// endState — what the store holds once the trace is over. Counts and
+// substrings, not rows: a trace is about who claimed what, and a payload
+// substring is enough to catch a request landing under the wrong words.
+type endState struct {
+	reminders []reminderWant
+	factKeys  []string
+	notes     int
+	tasks     []string
+}
+
+type reminderWant struct {
+	payload string // substring of the stored payload
+	fireAt  string // "2006-01-02 15:04" in UTC, "" ⇒ not checked
+}
+
+// trace — a named conversation, its turns, and the end state.
+type trace struct {
+	name  string
+	skip  string // non-empty ⇒ t.Skip: today's behaviour is wrong, this names the fix
+	turns []turn
+	end   endState
+}
+
+// newDialogueHandler — the offline floor with the real cascade and a movable
+// clock: newClarifyHandler's wiring (stub date parser, real fact parser, tool
+// matcher) plus the router newRoutingClarifyHandler builds, and the `now`
+// pointer so a turn can carry a wait.
+func newDialogueHandler(t *testing.T) (*reactiveHandler, *store.Store, *time.Time) {
+	t.Helper()
+	h, st, now := newClarifyHandler(t)
+	// A minute no trace ever says, so "fires at the current clock" is a defect
+	// and never a coincidence (V-577, V-579). checkEnd refuses any reminder
+	// landing on it, and at 09:00 the row that answers "на 9" would trip that.
+	*now = time.Date(2026, 7, 31, 9, 17, 0, 0, time.UTC)
+	h.router = buildRouter(router.NewHashEmbedder(1024), h.matcher, 0.55, nil, nil)
+	h.recall = recallWiring{embedder: router.NewHashEmbedder(1024), memStore: memory.NewInMemoryStore()}
+	return h, st, now
+}
+
+// wantedQuestion builds the question a turn must be answered with, from the
+// same code the daemon asks through. A time question is built from the gap,
+// because she names the clock and asks about the part he left out (V-579).
+func wantedQuestion(tn turn, now time.Time) (string, bool) {
+	if tn.question == dialogue.SlotTime {
+		gap := tn.gap
+		if gap == whenComplete {
+			gap = whenNoHour
+		}
+		return whenQuestion(gap, tn.attempt, now, whenTakenLine(tn.took))
+	}
+	return clarifyQuestionFor(tn.question, tn.attempt)
+}
+
+// runTrace drives one trace through handleText and checks every turn, then the
+// end state. Every failure carries the decision trace so far, so a wrong
+// claimant reads differently from wrong copy.
+func runTrace(t *testing.T, tr trace) {
+	t.Helper()
+	// MAVEN_DIALOGUE_NO_SKIP=1 runs the rows that fail today. That is how
+	// whoever lands V-560, V-561 or V-562 sees their row go green before
+	// deleting its skip, and it is also the check that a skip is still earned:
+	// a row that passes with the skip in place is a fix nobody noticed.
+	if tr.skip != "" && os.Getenv("MAVEN_DIALOGUE_NO_SKIP") == "" {
+		t.Skip(tr.skip)
+	}
+	ctx := context.Background()
+	h, st, now := newDialogueHandler(t)
+	const conversation = "web"
+	id := dialogueIDFor(sourceText, conversation)
+
+	var claims []claim
+	var previous string
+	fail := func(turnIdx int, format string, args ...any) {
+		t.Helper()
+		lines := make([]string, 0, len(claims))
+		for _, c := range claims {
+			lines = append(lines, "    "+c.String())
+		}
+		t.Fatalf("turn %d: "+format+"\n  who claimed what:\n%s",
+			append([]any{turnIdx}, append(args, strings.Join(lines, "\n"))...)...)
+	}
+
+	for i, tn := range tr.turns {
+		if tn.wait > 0 {
+			*now = now.Add(tn.wait)
+		}
+		var logged bytes.Buffer
+		prev := log.Writer()
+		log.SetOutput(&logged)
+		reply := h.handleText(ctx, conversation, tn.say)
+		log.SetOutput(prev)
+		claims = append(claims, claimsOf(tn.say, logged.String()))
+
+		body := reply
+		if tn.expired {
+			if !isClarifyExpired(reply) {
+				fail(i, "reply %q must open with the expiry notice", reply)
+			}
+			body = trimClarifyExpired(reply)
+			// The notice is glued in front of this turn's reply, and both halves
+			// have to survive: the words he just said are routed fresh, and
+			// answering only "I let the old one go" drops them.
+			if body == "" {
+				fail(i, "the notice was the whole reply; the fresh words were never answered")
+			}
+		} else if isClarifyExpired(reply) {
+			fail(i, "reply %q announced an expiry nothing asked for", reply)
+		}
+		if tn.question != "" {
+			want, ok := wantedQuestion(tn, h.now())
+			if !ok {
+				fail(i, "no question exists for slot %s attempt %d", tn.question, tn.attempt)
+			}
+			if body != want {
+				fail(i, "reply %q, want the %s question worded for attempt %d, %q", body, tn.question, tn.attempt, want)
+			}
+		}
+		if tn.noQuestion && isAnyClarifyQuestion(body) {
+			fail(i, "reply %q is another question; this turn is not something to ask about", body)
+		}
+		for _, want := range tn.contains {
+			if !strings.Contains(body, want) {
+				fail(i, "reply %q does not carry %q", body, want)
+			}
+		}
+		for _, unwanted := range tn.notContain {
+			if strings.Contains(body, unwanted) {
+				fail(i, "reply %q carries %q and must not", body, unwanted)
+			}
+		}
+		if tn.differs && reply == previous {
+			fail(i, "reply %q is byte-identical to the one before it, and his turn between them answered part of the gap", reply)
+		}
+		previous = reply
+		checkParked(t, fail, i, h.clarifyStore.Get(id, h.now()), tn.parked)
+	}
+	checkEnd(t, ctx, st, h, tr.end, claims)
+}
+
+// isAnyClarifyQuestion — is this reply one of her clarify questions, at any
+// attempt wording? Reads the templates rather than a list of its own.
+func isAnyClarifyQuestion(reply string) bool {
+	for _, variants := range clarifyQuestionVariants {
+		for _, v := range variants {
+			// HasSuffix, not equality: a question about the time opens with the
+			// clock she is reasoning from (V-579).
+			if strings.HasSuffix(reply, v) {
+				return true
+			}
+		}
+	}
+	// The two questions with no deck behind them, asked when the hour is said
+	// and its half of the day or its day is not.
+	return strings.HasSuffix(reply, "утра или вечера?") || strings.HasSuffix(reply, "В какой день?")
+}
+
+func checkParked(t *testing.T, fail func(int, string, ...any), i int, got *dialogue.PendingQuestion, want *parkedWant) {
+	t.Helper()
+	if want == nil {
+		if got != nil {
+			fail(i, "a question about %v is still armed and nothing should be: %+v", got.Missing, got.Slots)
+		}
+		return
+	}
+	if got == nil {
+		fail(i, "nothing is armed, want a question about %s (attempt %d)", want.slot, want.attempt)
+		return
+	}
+	if len(got.Missing) != 1 || got.Missing[0] != want.slot {
+		fail(i, "armed question is about %v, want %s", got.Missing, want.slot)
+	}
+	if got.Attempts != want.attempt {
+		fail(i, "armed question is on attempt %d, want %d — a retry spent on something that was never an answer is the V-554 shape", got.Attempts, want.attempt)
+	}
+	if want.carries != "" && !strings.Contains(got.Utterance, want.carries) {
+		fail(i, "the parked request no longer carries %q: %q", want.carries, got.Utterance)
+	}
+}
+
+func checkEnd(t *testing.T, ctx context.Context, st *store.Store, h *reactiveHandler, want endState, claims []claim) {
+	t.Helper()
+	lines := make([]string, 0, len(claims))
+	for _, c := range claims {
+		lines = append(lines, "    "+c.String())
+	}
+	trace := "\n  who claimed what:\n" + strings.Join(lines, "\n")
+
+	reminders, err := st.DueReminders(ctx, h.now().Add(14*24*time.Hour))
+	if err != nil {
+		t.Fatalf("DueReminders: %v", err)
+	}
+	if len(reminders) != len(want.reminders) {
+		t.Fatalf("end state: %d reminder(s), want %d: %+v%s", len(reminders), len(want.reminders), reminders, trace)
+	}
+	// No trace may leave a reminder at the current clock, whatever else it
+	// asserts (V-577, V-579). Twice on the box a sentence naming a day and no
+	// hour was completed from time.Now(): "что у меня сегодня?" became 01:28 and
+	// "на завтра" became 01:38. Neither minute was ever spoken, and a row that
+	// only checked the payload would have passed both.
+	for _, r := range reminders {
+		if r.FireTs.In(h.now().Location()).Format("15:04") == h.now().Format("15:04") {
+			t.Fatalf("end state: reminder %q fires at %s, which is the clock — a time slot naming no hour is asked about, never filled from now()%s",
+				r.Payload, r.FireTs.Format("15:04"), trace)
+		}
+	}
+	for i, w := range want.reminders {
+		if !strings.Contains(reminders[i].Payload, w.payload) {
+			t.Fatalf("end state: reminder %d payload %q does not carry %q%s", i, reminders[i].Payload, w.payload, trace)
+		}
+		if w.fireAt != "" {
+			if got := reminders[i].FireTs.UTC().Format("2006-01-02 15:04"); got != w.fireAt {
+				t.Fatalf("end state: reminder %d fires at %s, want %s%s", i, got, w.fireAt, trace)
+			}
+		}
+	}
+
+	facts, err := st.RecentFacts(ctx, 20)
+	if err != nil {
+		t.Fatalf("RecentFacts: %v", err)
+	}
+	if len(facts) != len(want.factKeys) {
+		t.Fatalf("end state: %d fact(s), want %d: %+v%s", len(facts), len(want.factKeys), facts, trace)
+	}
+	for i, key := range want.factKeys {
+		if facts[i].Key != key {
+			t.Fatalf("end state: fact %d is %q, want %q%s", i, facts[i].Key, key, trace)
+		}
+	}
+
+	notes, err := st.RecentNotes(ctx, 20)
+	if err != nil {
+		t.Fatalf("RecentNotes: %v", err)
+	}
+	if len(notes) != want.notes {
+		t.Fatalf("end state: %d note(s), want %d%s", len(notes), want.notes, trace)
+	}
+
+	tasks, err := st.ListTasks(ctx, store.TaskOpen)
+	if err != nil {
+		t.Fatalf("ListTasks: %v", err)
+	}
+	if len(tasks) != len(want.tasks) {
+		t.Fatalf("end state: %d open task(s), want %d: %+v%s", len(tasks), len(want.tasks), tasks, trace)
+	}
+	for i, text := range want.tasks {
+		if !strings.Contains(tasks[i].Text, text) {
+			t.Fatalf("end state: task %d is %q, want it to carry %q%s", i, tasks[i].Text, text, trace)
+		}
+	}
+}
+
+func TestDialogueTraces(t *testing.T) {
+	for _, tr := range dialogueTraces() {
+		tr := tr
+		t.Run(tr.name, func(t *testing.T) { runTrace(t, tr) })
+	}
+}
+
+// dialogueTraces — the fixture. Order is the order the shapes were found, not a
+// dependency: each trace builds its own handler and store.
+func dialogueTraces() []trace {
+	return []trace{
+		// The plain two-turn shape, and the one every other row is a deviation
+		// from: she asks for the time, he gives it, the reminder lands with the
+		// subject he said in the FIRST turn.
+		{
+			// Three turns since V-579, not two. An hour with no day named is
+			// not an answer she can act on: 11:00 today has passed as often as
+			// not, and picking one for him is the invention the whole rule is
+			// against. So she says the clock she is reasoning from and asks
+			// which day.
+			name: "reminder completed over three turns",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "маме"}},
+				{say: "в 11:00", question: dialogue.SlotTime, attempt: 2, gap: whenNoDay, took: "в 11:00",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 2, carries: "маме"}},
+				{say: "сегодня", contains: []string{"11:00"}, notContain: []string{"?"}},
+			},
+			end: endState{reminders: []reminderWant{{payload: "позвонить маме", fireAt: "2026-07-31 11:00"}}},
+		},
+		// The same shape on the fact path, where the answer carries both halves
+		// of what was missing — the key and the value — in one breath.
+		{
+			name: "fact completed over two turns",
+			turns: []turn{
+				{say: "запиши", question: dialogue.SlotKey, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotKey, attempt: 1}},
+				// His words back, not the key the parser filed them under
+				// (V-592). "water" is machine vocabulary and he never said it.
+				{say: "пил воду", contains: []string{"пил воду"}},
+			},
+			end: endState{factKeys: []string{"water"}},
+		},
+		// An answer past the TTL is a new request, not an answer (V-385). She
+		// says the old one is gone and routes the words fresh. A bare time on
+		// its own carries no request, so the fresh routing lands on the canned
+		// reply — the point of the row is that NOTHING is created: a reminder
+		// here would fire with the subject of a request she had already let go.
+		{
+			name: "answer arrives after the TTL",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "в 11:00", wait: clarifyTTL + time.Second, expired: true},
+			},
+			end: endState{},
+		},
+		// Three questions is the budget, and running out is SPOKEN: a mute
+		// give-up reads as "done" and he would wait for a reminder that was
+		// never set. The wording changes with the attempt (V-457).
+		{
+			name: "three unclear answers then the give-up line",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "ну не знаю", question: dialogue.SlotTime, attempt: 2,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 2}},
+				{say: "ну не знаю", question: dialogue.SlotTime, attempt: 3,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 3}},
+				{say: "ну не знаю", contains: []string{clarifyGaveUp}, noQuestion: true},
+			},
+			end: endState{},
+		},
+		// A correction points at the previous ACTED turn (repair.go): she redoes
+		// it under the intent he names and says so out loud, because a
+		// correction he cannot see is indistinguishable from one that was
+		// dropped. The task she filed first stays filed — repair redoes, it does
+		// not retract, and V-455 decided that deliberately.
+		//
+		// The corrected-to intent has to differ from the one she used, or repair
+		// declines: teaching the classifier the label it already produced is
+		// worse than doing nothing.
+		{
+			name: "correction of the previous turn",
+			turns: []turn{
+				{say: "добавь в задачи купить молоко", contains: []string{"купить молоко"}},
+				{say: "нет, это был вопрос", contains: []string{"поняла, это вопрос"}},
+			},
+			end: endState{tasks: []string{"купить молоко"}},
+		},
+		// He walks away from his own request: a question is parked, the next
+		// utterance is an unrelated request of its own, and nothing follows.
+		// V-554's fix is what makes this row pass — the question steps aside
+		// rather than scoring "добавь в задачи" as the time. The reminder is
+		// dropped in silence and that is the decision: if he meant it he says it
+		// again, and a question left armed eats the turn after next.
+		{
+			name: "abandoned flow: parked, then an unrelated request",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "добавь в задачи купить молоко", contains: []string{"купить молоко"}},
+				{say: "спасибо"},
+			},
+			end: endState{tasks: []string{"купить молоко"}},
+		},
+
+		// V-577 shape 1, the worst of the nine claimants measured on 2026-08-06.
+		// Every token of "что у меня сегодня?" is frame — an interrogative, a
+		// preposition, a particle and a day word — so the role classifier never
+		// looked at the route, the parked reminder read "сегодня" as its time,
+		// and the hour came from the clock. He got a reminder he never asked for
+		// at a minute he never said, and his question was answered nowhere.
+		//
+		// Two claims: the calendar answers, and nothing is written. The flow
+		// survives underneath, because a question of his own is not a request to
+		// abandon the one he was making.
+		{
+			name: "an agenda question mid-flow is answered, not eaten",
+			turns: []turn{
+				{say: "напомни забрать посылку", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "посылку"}},
+				{say: "что у меня сегодня?", contains: []string{"31.07.2026"},
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "посылку"}},
+			},
+			end: endState{},
+		},
+		// V-577 shape 2. He states something in the middle of the flow. It is
+		// neither a slot value nor a cancel, and it was scored as a failed
+		// answer and dropped in silence: alone the same sentence is stored.
+		// Silence is the one option that is wrong, so it is stored, no retry is
+		// spent, and the question comes back on the end of the same reply.
+		//
+		// The words are a fact and not the owner's note, because the fact parser
+		// is deterministic and the offline floor marks every classifier route
+		// Clarify. The row below carries his own sentence and needs the model.
+		//
+		// What this floor can prove is the arbitration: no retry is spent, the
+		// flow survives on the same attempt, and the words are answered as
+		// themselves with the question coming back after them. Whether the fact
+		// is then WRITTEN is the routing engine's business — the hash embedder
+		// is unsure of every sentence it sees, and an unsure fact has never been
+		// stored.
+		{
+			name: "a fact stated mid-flow steps aside without spending a retry",
+			turns: []turn{
+				{say: "напомни позвонить врачу", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "врачу"}},
+				// Nothing she says about it may be a word he did not say
+				// (V-592). On the box this sentence came back as "Проверила, что
+				// ты выпел стакан воды": a non-word for the verb, a glass copied
+				// out of the example in ReplySystemPrompt, and a claim to have
+				// checked something. The store held key=water value="drank"
+				// throughout, so all of it was generated from two tokens.
+				//
+				// The positive half of the contract — the confirmation IS his
+				// sentence — is asserted by "fact completed over two turns"
+				// above. It cannot be asserted here: the hash embedder marks
+				// this route Clarify, and an unsure fact is answered with the
+				// canned line rather than a confirmation of anything.
+				{say: "я выпил воды", contains: []string{"напоминание?"},
+					notContain: []string{"стакан", "выпел", "Проверила", "water"},
+					parked:     &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "врачу"}},
+			},
+			end: endState{},
+		},
+		// V-593: two asks about the same half of the day, with a turn between
+		// them that answered the DAY. Asking again is right and asking in the
+		// same bytes is not — from his side it is indistinguishable from not
+		// having been heard, which is what the whole V-558 family is about.
+		//
+		// The clock still opens every ask (the owner's rule, V-579); the
+		// acknowledgement goes after it and before the question.
+		{
+			name: "a re-ask names what the answer before it gave her",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "на 9", question: dialogue.SlotTime, attempt: 2, gap: whenAmbiguousHour, took: "на 9",
+					contains: []string{"Сейчас "},
+					parked:   &parkedWant{slot: dialogue.SlotTime, attempt: 2}},
+				{say: "на завтра", question: dialogue.SlotTime, attempt: 3, gap: whenAmbiguousHour, took: "на завтра",
+					contains: []string{"Сейчас ", "завтра"}, differs: true,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 3}},
+			},
+			end: endState{},
+		},
+		// V-579 turn 3: the preposition decided whether the hour was read. "в 9"
+		// set the reminder and "на 9" was not read at all, on the same build and
+		// with the same cardinal.
+		{
+			// It is read, and being read is not the same as being enough: nine is
+			// either half of the day, so she asks which and then which day
+			// (V-579). Both answers are frame words and neither carries an hour
+			// of its own, so this row is also the proof that an answer is read
+			// against the whole request rather than alone.
+			name: "на 9 answers the time question like в 9",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "на 9", question: dialogue.SlotTime, attempt: 2, gap: whenAmbiguousHour, took: "на 9",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 2}},
+				{say: "утра", question: dialogue.SlotTime, attempt: 3, gap: whenNoDay, took: "утра",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 3}},
+				{say: "завтра", contains: []string{"09:00"}},
+			},
+			end: endState{reminders: []reminderWant{{payload: "позвонить маме", fireAt: "2026-08-01 09:00"}}},
+		},
+		// The owner's own four, ruled 2026-08-06 (V-579). A reminder commits
+		// when what, what time and what day are all answered, and every ask
+		// states the clock she is reasoning from.
+		{
+			name: "his first example: a bare 3 is asked about",
+			turns: []turn{
+				{say: "напомни завтра в 3 заказать цветы",
+					question: dialogue.SlotTime, attempt: 1, gap: whenAmbiguousHour,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "цветы"}},
+			},
+			end: endState{},
+		},
+		{
+			// The hour is unambiguous and the day is still missing, so she asks.
+			// Today being a valid reading is not the same as him saying it.
+			name: "his second example: nine in the evening of which day",
+			turns: []turn{
+				{say: "напомни в 9 вечера разгрузить стиралку",
+					question: dialogue.SlotTime, attempt: 1, gap: whenNoDay,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "стиралку"}},
+				{say: "завтра", contains: []string{"21:00"}},
+			},
+			end: endState{reminders: []reminderWant{{payload: "стиралку", fireAt: "2026-08-01 21:00"}}},
+		},
+		{
+			// All three answered in one breath, so she does not ask at all.
+			name: "his third example: a full time commits",
+			turns: []turn{
+				{say: "напомни завтра в 15:00 заказать цветы", notContain: []string{"?"}},
+			},
+			end: endState{reminders: []reminderWant{{payload: "цветы", fireAt: "2026-08-01 15:00"}}},
+		},
+		{
+			// An interval is one instant, so it answers the hour and the day
+			// together. Confirmed by the owner: "через час is fine as is".
+			name: "an interval commits without a question",
+			turns: []turn{
+				{say: "напомни через час позвонить маме", notContain: []string{"?"}},
+			},
+			end: endState{reminders: []reminderWant{{payload: "маме", fireAt: "2026-07-31 10:17"}}},
+		},
+		// V-579 turn 4: he named a day and no hour, and got the day at the
+		// current minute. She has to ask instead, and the global check in
+		// checkEnd refuses the invented minute for every row at once.
+		{
+			name: "a day with no hour is asked about, not taken from the clock",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "на завтра", question: dialogue.SlotTime, attempt: 2,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 2}},
+			},
+			end: endState{},
+		},
+
+		// ---- rows below carry the CORRECT expectation and fail today ----
+
+		// The owner's own sentence from V-577 shape 2, in his words. It needs
+		// an engine that can route it: the hash embedder marks it note with
+		// Clarify set, and a route she is not sure of is not evidence that he
+		// stated anything. The row above is the same contract in words the
+		// floor's deterministic fact parser reads.
+		{
+			name: "a note stated mid-flow is stored, not dropped",
+			skip: "the offline floor cannot route «у меня новый ноутбук» confidently; needs the resident model",
+			turns: []turn{
+				{say: "напомни позвонить врачу", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "врачу"}},
+				{say: "у меня новый ноутбук",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "врачу"}},
+			},
+			end: endState{notes: 1},
+		},
+
+		// The owner's target transcript, V-561. He asks for a reminder, she asks
+		// when, he asks something else entirely, and then comes back to her
+		// question. On the box this created a reminder at 00:12 and never
+		// answered Rome; on the offline floor the side question is recognised as
+		// its own request and the flow is dropped instead, so the wrong reminder
+		// is not made and the right one is not either.
+		//
+		// Both are the same defect: there is no suspend and resume. The correct
+		// shape is the middle turn answered on its own and the parked question
+		// still standing, on the same attempt — a side query is not a failed
+		// answer and must not spend a retry.
+		//
+		// The skip came off with V-579. What held it was the parser, not the
+		// arbitration: neither the stub nor the production one read "на 9",
+		// because only "в" framed a spoken hour, and "на завтра" was completed
+		// from the clock.
+		//
+		// Turn 3 now closes the flow, where the transcript has one more exchange
+		// in it. That is the 12-hour question — the owner's turn 4 answers "на
+		// 9" with "сейчас 15:23, на 9 сегодня вечером?" — and it is a decision of
+		// its own, not one to invent here. Nine o'clock is read as nine and, at
+		// 09:17, as tomorrow's, which is where the transcript ends up anyway.
+		// Turn 4 then has nothing to answer and must not write anything.
+		{
+			name: "the owner's transcript from V-561",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "какая сейчас погода в Риме?",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "маме"}},
+				// His words, unchanged. What changed under V-579 is that "на 9"
+				// is a question and not a commit: nine could be either half of
+				// the day, so she says the clock she is reading from and asks.
+				// "на завтра." then answers the day and leaves the half open, so
+				// she asks that one again.
+				// Each ask names what the turn before it gave her (V-593). The
+				// two asks about the half of the day are the same question and
+				// must not be the same sentence: he answered between them, and a
+				// reply with no trace of that reads as not having been heard.
+				{say: "а, да, прости - на 9.", question: dialogue.SlotTime, attempt: 2, gap: whenAmbiguousHour, took: "а, да, прости - на 9.",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 2}},
+				{say: "на завтра.", question: dialogue.SlotTime, attempt: 3, gap: whenAmbiguousHour, took: "на завтра.",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 3}},
+			},
+			end: endState{},
+		},
+		// The same shape said in words StubDateTimeParser reads. GREEN since
+		// V-561. Same three claims: Rome is answered, the question survives the
+		// side query on the same attempt, and the answer after it completes the
+		// reminder he actually asked for.
+		//
+		// It sits under the "fail today" header because the row above it still
+		// does. Do not re-skip it to tidy that up: this is the owner's
+		// acceptance test in the only words the offline floor can read.
+		{
+			name: "nested question: a parked question, then one of his own",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "какая сейчас погода в Риме?",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "маме"}},
+				{say: "в 11:00", question: dialogue.SlotTime, attempt: 2, gap: whenNoDay, took: "в 11:00",
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 2, carries: "маме"}},
+				{say: "сегодня", contains: []string{"11:00"}},
+			},
+			end: endState{reminders: []reminderWant{{payload: "позвонить маме", fireAt: "2026-07-31 11:00"}}},
+		},
+		// A cancel is one of the five turn roles V-560 names, and today it is
+		// none of them: "неважно" fills no slot and carries no request of its
+		// own, so it reads as a failed answer and spends a retry. Two turns
+		// later she is still asking about a reminder he called off.
+		//
+		// The row asserts what is knowable — nothing armed, nothing written, and
+		// not another question — rather than her wording for it, which is not
+		// written yet and is not this task's to invent.
+		{
+			name: "cancel: a parked question, then never mind",
+			skip: "V-560: a cancel is scored as a failed answer, not as a cancel",
+			turns: []turn{
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "неважно", noQuestion: true},
+			},
+			end: endState{},
+		},
+		// Order in runTurn is the whole arbitration (V-558), and this is what it
+		// costs: the clarify answer is checked at step 3 and the repair marker at
+		// step 4d, so while a question is parked no correction can be made. She
+		// scores "нет, это была заметка" as a bad time answer and asks again.
+		{
+			name: "correction while a question is parked",
+			skip: "V-560: clarify pre-empts the repair marker, so a correction cannot be spoken mid-flow",
+			turns: []turn{
+				{say: "добавь в задачи купить молоко", contains: []string{"купить молоко"}},
+				{say: "напомни позвонить маме", question: dialogue.SlotTime, attempt: 1,
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+				{say: "нет, это был вопрос", contains: []string{"поняла, это вопрос"},
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1, carries: "маме"}},
+			},
+			end: endState{tasks: []string{"купить молоко"}},
+		},
+		// A reminder said whole, in one breath, with the hour in it — and she
+		// asks when. ReminderGrammar (stage0.go) builds its slots by hand and
+		// never runs the extractor, so a stage-0 reminder carries no time
+		// whatever the sentence says, and the clarify gate reads the gap as
+		// real. It costs a turn on the commonest reminder shape there is.
+		//
+		// Hermetic despite the date parser: stage 0 calls no parser at all, so
+		// this fails the same way with or without python dateparser installed.
+		{
+			name: "a reminder said whole is not asked about",
+			skip: "V-562: a stage-0 decision never meets the extractor, so its slots are never validated",
+			turns: []turn{
+				{say: "напомни в 11:00 позвонить маме", contains: []string{"11:00"}, noQuestion: true},
+			},
+			end: endState{reminders: []reminderWant{{payload: "позвонить маме", fireAt: "2026-07-31 11:00"}}},
+		},
+		// The same gap on the repair path. A correction redoes the request
+		// through finishClarified, which goes straight to applyAction — it never
+		// passes the clarify gate — so a redo that lands short answers with the
+		// parse error V-557 removed from the routing path: "не поняла, на когда
+		// напомнить." She should ask, exactly as she does for a fresh reminder
+		// with no time.
+		{
+			name: "a correction that lands short asks rather than failing",
+			skip: "V-562: finishClarified skips the clarify gate, so a repaired decision is never checked for gaps",
+			turns: []turn{
+				{say: "добавь в задачи купить молоко", contains: []string{"купить молоко"}},
+				{say: "нет, это было напоминание", contains: []string{"поняла, это напоминание"},
+					parked: &parkedWant{slot: dialogue.SlotTime, attempt: 1}},
+			},
+			end: endState{tasks: []string{"купить молоко"}},
+		},
+	}
+}

cmd/mavend/ecosystem.go

@@ -12,6 +12,7 @@ import (
 	"log"
 	"net/http"
 	"net/url"
+	"strings"
 	"time"
 
 	hexisclient "github.com/kami/hexis/pkg/client"
@@ -42,28 +43,100 @@ const ecosystemAPIVersion = "v1"
 // anonymous HTTP client.
 const mavenRequester = "maven"
 
-// setEcosystemHeaders stamps the version, requester, auth and correlation
-// headers common to every outgoing ecosystem request. token may be empty,
-// which means the transport itself is trusted (loopback or unix socket).
+// ecosystemHTTP is the JSON transport every ecosystem client shares: one base
+// URL, one bearer token, and the header set the contract requires on each
+// request. Nexus and Praxis differ only in the service name and the version
+// header, so both embed this rather than repeating build, send and classify.
+type ecosystemHTTP struct {
+	service       string // "nexus", "praxis" — the name errors and traces carry
+	versionHeader string
+	baseURL       string
+	token         string
+	httpClient    *http.Client
+}
+
+func newEcosystemHTTP(service, versionHeader, baseURL string) ecosystemHTTP {
+	return ecosystemHTTP{
+		service:       service,
+		versionHeader: versionHeader,
+		baseURL:       baseURL,
+		httpClient:    &http.Client{Timeout: 10 * time.Second},
+	}
+}
+
+// setHeaders stamps the version, requester, auth and correlation headers common
+// to every outgoing ecosystem request. The token may be empty, which means the
+// transport itself is trusted (loopback or unix socket).
 //
-// The correlation ID is read from the context and never minted here. Minting
-// one per request sent the far side an ID that existed nowhere on this side,
-// and gave a single multi-hop action as many unrelated IDs as it made calls.
-// Callers that start an action assign the ID once (handleHexisAct,
+// The correlation ID is read from the request's own context and never minted
+// here. Minting one per request sent the far side an ID that existed nowhere on
+// this side, and gave a single multi-hop action as many unrelated IDs as it
+// made calls. Callers that start an action assign the ID once (handleHexisAct,
 // handlePraxisAct, resolveEntityReference) and every hop inherits it.
-func setEcosystemHeaders(req *http.Request, ctx context.Context, versionHeader, token string) {
+func (t *ecosystemHTTP) setHeaders(req *http.Request) {
 	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set(versionHeader, ecosystemAPIVersion)
+	req.Header.Set(t.versionHeader, ecosystemAPIVersion)
 	req.Header.Set("Accept", "application/json")
 	req.Header.Set("X-Requested-By", mavenRequester)
-	if token != "" {
-		req.Header.Set("Authorization", "Bearer "+token)
+	if t.token != "" {
+		req.Header.Set("Authorization", "Bearer "+t.token)
 	}
-	if id := correlationIDFromCtx(ctx); id != "" {
+	if id := correlationIDFromCtx(req.Context()); id != "" {
 		req.Header.Set("X-Correlation-ID", id)
 	}
 }
 
+// call sends one request and decodes the JSON answer into out, which may be nil
+// when the body carries nothing worth reading. op is the logical operation name
+// for errors and traces: the path carries the query string, and after entity
+// scoping that means an entity id in every log line built from the error, next
+// to a trace that redacts far less than that.
+//
+// Every failure is an *ecosystemError, including the transport and decode ones.
+// Some of these paths mutate remote state, and the question worth answering
+// afterwards is whether the call never left or was refused.
+func (t *ecosystemHTTP) call(ctx context.Context, method, op, path string, payload, out any) error {
+	var body io.Reader
+	if payload != nil {
+		data, err := json.Marshal(payload)
+		if err != nil {
+			return &ecosystemError{Service: t.service, Op: op, Err: err}
+		}
+		body = bytes.NewReader(data)
+	}
+	req, err := http.NewRequestWithContext(ctx, method, t.baseURL+path, body)
+	if err != nil {
+		return &ecosystemError{Service: t.service, Op: op, Err: err}
+	}
+	t.setHeaders(req)
+
+	resp, err := t.httpClient.Do(req)
+	if err != nil {
+		return &ecosystemError{Service: t.service, Op: op, Err: err}
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != 200 {
+		return httpError(t.service, op, resp.StatusCode)
+	}
+	if out == nil {
+		return nil
+	}
+	if err := json.NewDecoder(resp.Body).Decode(out); err != nil {
+		return &ecosystemError{Service: t.service, Op: op, Status: resp.StatusCode, Err: err}
+	}
+	return nil
+}
+
+// getJSON performs a GET and decodes the JSON body into out.
+func (t *ecosystemHTTP) getJSON(ctx context.Context, op, path string, out any) error {
+	return t.call(ctx, http.MethodGet, op, path, nil, out)
+}
+
+// postJSON posts a JSON payload and decodes the JSON answer into out.
+func (t *ecosystemHTTP) postJSON(ctx context.Context, op, path string, payload, out any) error {
+	return t.call(ctx, http.MethodPost, op, path, payload, out)
+}
+
 // ecosystemError is the typed failure every ecosystem client returns, so
 // callers can tell a transport failure from a refusal from a contract
 // mismatch without matching on message text. The distinction matters:
@@ -107,17 +180,60 @@ func httpError(service, op string, status int) *ecosystemError {
 	}
 }
 
+// hexisStatusTexts maps the http.StatusText spelling back to its code, for the
+// failure statuses a Hexis call can plausibly answer with. It is the inverse of
+// what the vendored client threw away.
+var hexisStatusTexts = func() map[string]int {
+	codes := []int{
+		http.StatusBadRequest, http.StatusUnauthorized, http.StatusForbidden,
+		http.StatusNotFound, http.StatusMethodNotAllowed, http.StatusNotAcceptable,
+		http.StatusRequestTimeout, http.StatusConflict, http.StatusGone,
+		http.StatusUnprocessableEntity, http.StatusUpgradeRequired,
+		http.StatusTooManyRequests, http.StatusInternalServerError,
+		http.StatusNotImplemented, http.StatusBadGateway,
+		http.StatusServiceUnavailable, http.StatusGatewayTimeout,
+	}
+	m := make(map[string]int, len(codes))
+	for _, c := range codes {
+		m[http.StatusText(c)] = c
+	}
+	return m
+}()
+
+// hexisError re-wraps an error from the vendored Hexis client as an
+// *ecosystemError, so a Hexis failure classifies the same way a Nexus or Praxis
+// one does and ecosystemGap can tell a refused credential from an outage.
+//
+// This is a boundary adapter and it is not the fix anyone would choose. The
+// Hexis client lives in another repository and returns
+// fmt.Errorf("%s: %s", http.StatusText(status), body) for every status at or
+// above 400, so the status text is the only signal that survives — the correct
+// fix is a typed error carrying the code, and Maven cannot land it unilaterally
+// (Vikunja #587, docs/plans/20-two-artifacts-and-neither-is-spring.md). Parsing
+// here is bounded: the message's first colon-delimited segment is the status
+// text verbatim, no status text contains a colon, and anything unrecognised —
+// "do request: ...", "create request: ..." — is a transport failure and is left
+// at status 0, which is exactly what Unreachable() means.
+func hexisError(op string, err error) error {
+	if err == nil {
+		return nil
+	}
+	var ee *ecosystemError
+	if errors.As(err, &ee) {
+		return err
+	}
+	head, _, _ := strings.Cut(err.Error(), ": ")
+	return &ecosystemError{
+		Service: "hexis", Op: op, Status: hexisStatusTexts[head], Err: err,
+	}
+}
+
 type nexusClient struct {
-	baseURL    string
-	token      string
-	httpClient *http.Client
+	ecosystemHTTP
 }
 
 func newNexusClient(url string) *nexusClient {
-	return &nexusClient{
-		baseURL:    url,
-		httpClient: &http.Client{Timeout: 10 * time.Second},
-	}
+	return &nexusClient{newEcosystemHTTP("nexus", "X-Nexus-Version", url)}
 }
 
 // withToken sets the bearer token sent on every request. Returns the client so
@@ -176,62 +292,26 @@ func (c *nexusClient) Resolve(ctx context.Context, query string, types []string)
 		body["types"] = types
 	}
 
-	data, _ := json.Marshal(body)
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, c.baseURL+"/api/v1/resolve", bytes.NewReader(data))
-	if err != nil {
-		return nil, fmt.Errorf("create request: %w", err)
-	}
-	setEcosystemHeaders(req, ctx, "X-Nexus-Version", c.token)
-
-	resp, err := c.httpClient.Do(req)
-	if err != nil {
-		return nil, &ecosystemError{Service: "nexus", Op: "resolve", Err: err}
-	}
-	defer resp.Body.Close()
-
-	bodyBytes, _ := io.ReadAll(resp.Body)
-	if resp.StatusCode != 200 {
-		return nil, httpError("nexus", "resolve", resp.StatusCode)
-	}
-
 	var result nexusResolveResult
-	if err := json.Unmarshal(bodyBytes, &result); err != nil {
-		return nil, &ecosystemError{Service: "nexus", Op: "resolve", Status: resp.StatusCode, Err: err}
+	if err := c.postJSON(ctx, "resolve", "/api/v1/resolve", body, &result); err != nil {
+		return nil, err
 	}
 	return &result, nil
 }
 
 func (c *nexusClient) Health(ctx context.Context) error {
-	req, err := http.NewRequestWithContext(ctx, http.MethodGet, c.baseURL+"/health", nil)
-	if err != nil {
-		return &ecosystemError{Service: "nexus", Op: "health", Err: err}
-	}
-	setEcosystemHeaders(req, ctx, "X-Nexus-Version", c.token)
-	resp, err := c.httpClient.Do(req)
-	if err != nil {
-		return &ecosystemError{Service: "nexus", Op: "health", Err: err}
-	}
-	resp.Body.Close()
-	if resp.StatusCode != 200 {
-		return httpError("nexus", "health", resp.StatusCode)
-	}
-	return nil
+	return c.getJSON(ctx, "health", "/health", nil)
 }
 
 // praxisClient talks to the Praxis HTTP tools API. Maven must not open Praxis's
 // SQLite store directly (ecosystem invariant: no component reads another's DB),
 // so attention/changes/lifecycle all go over this HTTP contract against praxisd.
 type praxisClient struct {
-	baseURL    string
-	token      string
-	httpClient *http.Client
+	ecosystemHTTP
 }
 
 func newPraxisClient(url string) *praxisClient {
-	return &praxisClient{
-		baseURL:    url,
-		httpClient: &http.Client{Timeout: 10 * time.Second},
-	}
+	return &praxisClient{newEcosystemHTTP("praxis", "X-Praxis-Version", url)}
 }
 
 func (c *praxisClient) withToken(token string) *praxisClient {
@@ -239,30 +319,6 @@ func (c *praxisClient) withToken(token string) *praxisClient {
 	return c
 }
 
-// getJSON performs a GET and decodes the JSON body into out. op is the logical
-// operation name for errors and traces: the path carries the query string, and
-// after entity scoping that means an entity id in every log line built from the
-// error, next to a trace that redacts far less than that.
-func (c *praxisClient) getJSON(ctx context.Context, op, path string, out any) error {
-	req, err := http.NewRequestWithContext(ctx, http.MethodGet, c.baseURL+path, nil)
-	if err != nil {
-		return err
-	}
-	setEcosystemHeaders(req, ctx, "X-Praxis-Version", c.token)
-	resp, err := c.httpClient.Do(req)
-	if err != nil {
-		return &ecosystemError{Service: "praxis", Op: op, Err: err}
-	}
-	defer resp.Body.Close()
-	if resp.StatusCode != 200 {
-		return httpError("praxis", op, resp.StatusCode)
-	}
-	if err := json.NewDecoder(resp.Body).Decode(out); err != nil {
-		return &ecosystemError{Service: "praxis", Op: op, Status: resp.StatusCode, Err: err}
-	}
-	return nil
-}
-
 // praxisAttention — an attention response in either of the two shapes Praxis
 // may send (Vikunja #540).
 //
@@ -376,31 +432,14 @@ type praxisItem struct {
 // postItemAction posts {"item_id": id} to a Praxis tools lifecycle endpoint
 // and decodes the resulting item. Shared by Surface/Acknowledge/Resolve/Ignore.
 func (c *praxisClient) postItemAction(ctx context.Context, op, path, itemID string) (*praxisItem, error) {
-	return c.postJSON(ctx, op, path, map[string]any{"item_id": itemID})
+	return c.postItem(ctx, op, path, map[string]any{"item_id": itemID})
 }
 
-// postJSON posts a body to a Praxis lifecycle endpoint and decodes the item.
-// Every failure is a *ecosystemError, including the transport and decode ones:
-// these are the paths that mutate remote state, and the question worth
-// answering afterwards is whether the call never left or was refused.
-func (c *praxisClient) postJSON(ctx context.Context, op, path string, payload map[string]any) (*praxisItem, error) {
-	body, _ := json.Marshal(payload)
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, c.baseURL+path, bytes.NewReader(body))
-	if err != nil {
-		return nil, &ecosystemError{Service: "praxis", Op: op, Err: err}
-	}
-	setEcosystemHeaders(req, ctx, "X-Praxis-Version", c.token)
-	resp, err := c.httpClient.Do(req)
-	if err != nil {
-		return nil, &ecosystemError{Service: "praxis", Op: op, Err: err}
-	}
-	defer resp.Body.Close()
-	if resp.StatusCode != 200 {
-		return nil, httpError("praxis", op, resp.StatusCode)
-	}
+// postItem posts a body to a Praxis lifecycle endpoint and decodes the item.
+func (c *praxisClient) postItem(ctx context.Context, op, path string, payload map[string]any) (*praxisItem, error) {
 	var out praxisItem
-	if err := json.NewDecoder(resp.Body).Decode(&out); err != nil {
-		return nil, &ecosystemError{Service: "praxis", Op: op, Status: resp.StatusCode, Err: err}
+	if err := c.postJSON(ctx, op, path, payload, &out); err != nil {
+		return nil, err
 	}
 	return &out, nil
 }
@@ -425,7 +464,7 @@ func (c *praxisClient) Ignore(ctx context.Context, itemID string) (*praxisItem,
 }
 
 func (c *praxisClient) Pin(ctx context.Context, itemID string, pinned bool) (*praxisItem, error) {
-	return c.postJSON(ctx, "pin", "/api/v1/tools/pin", map[string]any{"item_id": itemID, "pinned": pinned})
+	return c.postItem(ctx, "pin", "/api/v1/tools/pin", map[string]any{"item_id": itemID, "pinned": pinned})
 }
 
 func (c *praxisClient) GetItem(ctx context.Context, itemID string) (*praxisItem, error) {
@@ -540,6 +579,7 @@ func (w *ecosystemWiring) discoverCapabilities(ctx context.Context, entityID str
 	}
 	caps, err := w.hexis.Capabilities(ctx, entityID)
 	if err != nil {
+		err = hexisError("capabilities", err)
 		log.Printf("ecosystem: hexis capabilities error: %v", err)
 		return nil, err
 	}
@@ -567,7 +607,10 @@ func (w *ecosystemWiring) executeCapability(ctx context.Context, capabilityID, t
 
 	exec, err := w.hexis.Execute(ctx, req)
 	if err != nil {
-		return correlationID, fmt.Errorf("execute: %w", err)
+		// A classified dependency failure. The two returns below are NOT: an
+		// execution that ran and failed is the command failing, not Hexis
+		// degrading, and it keeps its plain error so the caller says so.
+		return correlationID, hexisError("execute", err)
 	}
 	if exec.Status == "succeeded" {
 		return correlationID, nil

cmd/mavend/ecosystem_acts.go

@@ -22,13 +22,25 @@ import (
 // already knows which one it was talking to — it records the same name in the
 // trace (Vikunja #521).
 const (
-	serviceNexus = "Nexus"
-	serviceHexis = "Hexis"
+	serviceNexus  = "Nexus"
+	servicePraxis = "Praxis"
+	serviceHexis  = "Hexis"
 )
 
 // serviceVars — the one-key map the eco_down and eco_denied lines take.
 func serviceVars(name string) map[string]string { return map[string]string{"name": name} }
 
+// ecosystemGap names the service that failed. A rejected credential gets its
+// own line, because a wrong token looks exactly like an outage to him and
+// "try again" is advice that will never work. Every degrade path reads through
+// here, so all of them name the service and none of them guesses instead.
+func ecosystemGap(service string, err error) string {
+	if unauthorizedEcosystemError(err) {
+		return phraser.A(phraser.EcoDenied, serviceVars(service))
+	}
+	return phraser.A(phraser.EcoDown, serviceVars(service))
+}
+
 // praxisCapability is one arm of the Praxis act dispatch. This is an interface
 // rather than a map[string]func because each arm carries its own state: the
 // verb aliases it answers to, the trace name it records, and its own reply
@@ -127,10 +139,16 @@ func (h *reactiveHandler) handlePraxisAct(ctx context.Context, dec router.Decisi
 // praxisItemAction is the shared shape of the item-lifecycle capabilities: take
 // an item id from the value slot, call one Praxis endpoint, trace the result.
 type praxisItemAction struct {
-	verbs   []string
-	ask     string // reply when no item id was given
-	op      string // trace + log name of the operation
-	failure string // reply when the Praxis call errors
+	verbs []string
+	ask   string // reply when no item id was given
+	op    string // trace + log name of the operation
+	// failure is the first half of the reply when the Praxis call errors: which
+	// operation did not happen. ecosystemGap supplies the second half, which
+	// names Praxis and splits a refused token from an outage — those two used to
+	// produce the identical sentence and neither said "Praxis" (Vikunja #588).
+	// The verb is kept alongside the service name because the trace is the only
+	// other place it exists, and he is not reading the trace.
+	failure string
 	success string
 	call    func(ctx context.Context, px *praxisClient, id string) error
 }
@@ -147,7 +165,7 @@ func (a praxisItemAction) handle(ctx context.Context, h *reactiveHandler, px *pr
 		log.Printf("ecosystem: praxis %s %s: %v", a.op, id, err)
 		h.recordEcosystemTrace(ctx, "praxis", a.op, traceStatusForError(err), started,
 			mergeFields(traceErrorFields(err), map[string]any{"item_id": id}))
-		return a.failure
+		return a.failure + " " + ecosystemGap(servicePraxis, err)
 	}
 	h.recordPraxisTrace(ctx, a.op, started, map[string]any{"item_id": id})
 	return a.success
@@ -199,16 +217,10 @@ func (listAttentionCapability) handle(ctx context.Context, h *reactiveHandler, p
 		}
 		parts = append(parts, s)
 
-		// Speaking an item surfaces it, it does not acknowledge it
-		// (ECOSYSTEM-SPEC.md §2.3: surfaced != acknowledged). Best-effort:
-		// a failed surface call must not block delivering the digest.
-		if id, ok := item["id"].(string); ok && id != "" {
-			// Recorded in the order she says them, and only for items she could
-			// say: an item skipped above has no position in what he heard (#516).
+		// Recorded in the order she says them, and only for items she could
+		// say: an item skipped above has no position in what he heard (#516).
+		if id := surfaceSpoken(ctx, px, item); id != "" {
 			spoken = append(spoken, id)
-			if _, err := px.Surface(ctx, id); err != nil {
-				log.Printf("ecosystem: praxis surface %s: %v", id, err)
-			}
 		}
 	}
 	h.rememberSurfaced(spoken)
@@ -298,12 +310,7 @@ func (entityAttentionCapability) handle(ctx context.Context, h *reactiveHandler,
 		// trace gets. A trace that stores a rune count next to a log line
 		// storing the runes is not redacted at all.
 		log.Printf("ecosystem: entity attention resolve %s: %v", redactSubject(subject), err)
-		h.recordEcosystemTrace(ctx, "nexus", "resolve", traceStatusForError(err), started,
-			mergeFields(traceErrorFields(err), map[string]any{"subject": redactSubject(subject)}))
-		if unauthorizedEcosystemError(err) {
-			return phraser.A(phraser.EcoDenied, serviceVars(serviceNexus))
-		}
-		return phraser.A(phraser.EcoDown, serviceVars(serviceNexus))
+		return h.nexusResolveFailed(ctx, subject, started, err)
 	}
 	if len(ambiguous) > 0 {
 		return phraser.A(phraser.EcoAmbiguous, map[string]string{"items": strings.Join(ambiguous, ", ")})
@@ -339,19 +346,24 @@ func (entityAttentionCapability) handle(ctx context.Context, h *reactiveHandler,
 	})
 
 	var parts []string
+	var spoken []string
 	for _, item := range items {
 		title, _ := item["title"].(string)
 		if title == "" {
 			continue
 		}
 		parts = append(parts, title)
-		// Same surfaced != acknowledged rule as the unscoped digest.
-		if id, ok := item["id"].(string); ok && id != "" {
-			if _, err := px.Surface(ctx, id); err != nil {
-				log.Printf("ecosystem: praxis surface %s: %v", id, err)
-			}
+		if id := surfaceSpoken(ctx, px, item); id != "" {
+			spoken = append(spoken, id)
 		}
 	}
+	// The scoped digest is a list she read out, so it replaces the positional
+	// memory exactly as the unscoped one does. It used to surface these items
+	// and remember none of them, which left the previous digest live: "отметь
+	// второй как сделанное" then indexed into a list he had not just heard and
+	// transitioned somebody else's item (docs/ecosystem.md — a wrong guess here
+	// transitions the wrong item).
+	h.rememberSurfaced(spoken)
 	if known := h.localFactsForEntity(ctx, entityID); known != "" {
 		parts = append(parts, known)
 	}
@@ -366,6 +378,22 @@ func (entityAttentionCapability) handle(ctx context.Context, h *reactiveHandler,
 	return phraser.A(phraser.AttentionListEntity, map[string]string{"name": displayName, "items": strings.Join(parts, "; ")})
 }
 
+// surfaceSpoken marks an item she just read out as surfaced. Speaking an item
+// surfaces it, it does not acknowledge it (ECOSYSTEM-SPEC.md §2.3: surfaced !=
+// acknowledged), so this calls Surface and nothing else. Best effort: a failed
+// surface call must not block delivering the digest. Returns the item id, or ""
+// when the item carried none.
+func surfaceSpoken(ctx context.Context, px *praxisClient, item map[string]any) string {
+	id, _ := item["id"].(string)
+	if id == "" {
+		return ""
+	}
+	if _, err := px.Surface(ctx, id); err != nil {
+		log.Printf("ecosystem: praxis surface %s: %v", id, err)
+	}
+	return id
+}
+
 // scopedToEntity drops items that carry an entity_id other than the one asked
 // about, and reports whether the response can be trusted as scoped at all. An
 // item without an entity_id is kept only when at least one sibling carries the
@@ -474,6 +502,14 @@ func traceStatusForError(err error) string {
 	return traceFailed
 }
 
+// nexusResolveFailed records a resolve that failed and returns the named gap.
+// The subject is his words, so the trace keeps a rune count and not the runes.
+func (h *reactiveHandler) nexusResolveFailed(ctx context.Context, subject string, started time.Time, err error) string {
+	h.recordEcosystemTrace(ctx, "nexus", "resolve", traceStatusForError(err), started,
+		mergeFields(traceErrorFields(err), map[string]any{"subject": redactSubject(subject)}))
+	return ecosystemGap(serviceNexus, err)
+}
+
 // redactSubject reduces a user utterance to something safe to persist in a
 // trace: its length only. Traces are diagnostics, and his words are not
 // diagnostics — the correlation ID is what ties a trace to the turn.
@@ -532,26 +568,31 @@ func unauthorizedEcosystemError(err error) bool {
 	return errors.As(err, &ee) && ee.Unauthorized()
 }
 
+// isEcosystemError reports a failure that belongs to the service rather than to
+// what was asked of it: a call that never landed, or one the far side refused.
+// It separates "Hexis is down" from "the restart failed".
+func isEcosystemError(err error) bool {
+	var ee *ecosystemError
+	return errors.As(err, &ee)
+}
+
 // traceErrorFields describes an ecosystemError for a trace without leaking the
 // payload: the HTTP status and the failure class, nothing else.
 func traceErrorFields(err error) map[string]any {
-	fields := map[string]any{}
+	fields := map[string]any{"class": "error"}
 	var ee *ecosystemError
-	if errors.As(err, &ee) {
-		fields["http_status"] = ee.Status
-		switch {
-		case ee.Unauthorized():
-			fields["class"] = "unauthorized"
-		case ee.ContractMismatch():
-			fields["class"] = "contract_mismatch"
-		case ee.Unreachable():
-			fields["class"] = "unreachable"
-		default:
-			fields["class"] = "error"
-		}
+	if !errors.As(err, &ee) {
 		return fields
 	}
-	fields["class"] = "error"
+	fields["http_status"] = ee.Status
+	switch {
+	case ee.Unauthorized():
+		fields["class"] = "unauthorized"
+	case ee.ContractMismatch():
+		fields["class"] = "contract_mismatch"
+	case ee.Unreachable():
+		fields["class"] = "unreachable"
+	}
 	return fields
 }
 
@@ -636,16 +677,11 @@ func (h *reactiveHandler) handleHexisAct(ctx context.Context, dec router.Decisio
 	res := h.resolveEntityCandidates(ctx, entityReferences(dec))
 	subject, entityID, displayName, ambiguous, err := res.subject, res.entityID, res.displayName, res.ambiguous, res.err
 	if err != nil {
-		h.recordEcosystemTrace(ctx, "nexus", "resolve", traceStatusForError(err), started,
-			mergeFields(traceErrorFields(err), map[string]any{"subject": redactSubject(subject)}))
-		if unauthorizedEcosystemError(err) {
-			return phraser.A(phraser.EcoDenied, serviceVars(serviceNexus))
-		}
 		// A genuine Nexus dependency failure, not "no such entity" — stop here
 		// and report degradation rather than silently falling through to the
 		// local command executor (ECOSYSTEM-SPEC.md: services degrade
 		// independently, never a silent all-clear).
-		return phraser.A(phraser.EcoDown, serviceVars(serviceNexus))
+		return h.nexusResolveFailed(ctx, subject, started, err)
 	}
 	if len(ambiguous) > 0 {
 		h.recordEcosystemTrace(ctx, "nexus", "resolve", traceAmbig, started,
@@ -668,10 +704,7 @@ func (h *reactiveHandler) handleHexisAct(ctx context.Context, dec router.Decisio
 	if err != nil {
 		h.recordEcosystemTrace(ctx, "hexis", "capabilities", traceStatusForError(err), discovered,
 			mergeFields(traceErrorFields(err), map[string]any{"entity_id": entityID}))
-		if unauthorizedEcosystemError(err) {
-			return phraser.A(phraser.EcoDenied, serviceVars(serviceHexis))
-		}
-		return phraser.A(phraser.EcoDown, serviceVars(serviceHexis))
+		return ecosystemGap(serviceHexis, err)
 	}
 	h.recordEcosystemTrace(ctx, "hexis", "capabilities", traceOK, discovered,
 		map[string]any{"entity_id": entityID, "count": len(caps)})
@@ -745,6 +778,9 @@ func (h *reactiveHandler) handleHexisAct(ctx context.Context, dec router.Decisio
 			entityID:     entityID,
 			displayName:  displayName,
 			expiry:       h.now().Add(confirmTTL),
+			// This action's id, so the execution the confirm authorises is
+			// joined to the resolve and the discovery that proposed it.
+			correlationID: correlationIDFromCtx(ctx),
 		}
 		h.mu.Unlock()
 		h.recordEcosystemTrace(ctx, "hexis", "confirmation", tracePending, started,
@@ -769,6 +805,14 @@ func (h *reactiveHandler) execHexis(ctx context.Context, capID, capName, entityI
 			mergeFields(traceErrorFields(err), map[string]any{
 				"entity_id": entityID, "capability": capName, "causation_id": causationID,
 			}))
+		// Hexis never answering, or answering "no", is a gap in Hexis and is
+		// named as one — a refused token said "не получилось выполнить команду"
+		// here and sent him to debug a capability that was never reached
+		// (Vikunja #587). An execution that genuinely ran and failed is not an
+		// ecosystemError and keeps the command-level line.
+		if isEcosystemError(err) {
+			return ecosystemGap(serviceHexis, err)
+		}
 		return phraser.A(phraser.ActFailEntity, map[string]string{"name": displayName})
 	}
 	// One record per hop: the second write this used to make said the same

cmd/mavend/ecosystem_gap_test.go

@@ -0,0 +1,159 @@
+package main
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/phraser"
+)
+
+// A refused credential and an outage are different answers, and on the Hexis
+// path only one of them used to be said. These tests pin the difference at both
+// Hexis sites: the discovery hop and the execute hop (Vikunja #587). The Praxis
+// half of the same defect is in praxis_gap_test.go.
+//
+// unreachableURL is a port nothing listens on, which is what "the service is
+// down" looks like from inside a call: the connection is refused, no HTTP
+// answer is ever produced, and ecosystemError.Unreachable() is true.
+const unreachableURL = "http://127.0.0.1:1"
+
+func denied(service, reply string) bool {
+	return phraser.IsA(phraser.EcoDenied, serviceVars(service), reply)
+}
+
+func down(service, reply string) bool {
+	return phraser.IsA(phraser.EcoDown, serviceVars(service), reply)
+}
+
+// hexisGapHandler wires a handler whose Nexus resolves cleanly and whose Hexis
+// is the caller's to break. hexisURL is taken separately so a test can point it
+// at a dead port.
+func hexisGapHandler(t *testing.T, nexusURL, hexisURL string) *reactiveHandler {
+	t.Helper()
+	st := newTestStore(t)
+	now := time.Now()
+	return &reactiveHandler{
+		api:       ipc.NewStoreAPI(st),
+		dataStore: st,
+		now:       func() time.Time { return now },
+		ecosystem: stubEcosystem(nexusURL, hexisURL),
+	}
+}
+
+// TestHexisDiscovery401IsDeniedNotDown — the discovery hop.
+//
+// The vendored Hexis client returns a plain fmt.Errorf for every status at or
+// above 400, so errors.As for *ecosystemError never matched and every failure
+// fell through to the outage line. "Hexis is down" for a rejected token sends
+// him to inspect a service that is running fine.
+func TestHexisDiscovery401IsDeniedNotDown(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", muzickIndexer, "service"))
+	caps := fixtureHexisCapabilities(map[string]any{"id": "cap_restart", "name": "restart", "read_only": true})
+	hexis := newFakeHexis(t, caps, fixtureHexisExecuted("exec_1", "succeeded"))
+	h := hexisGapHandler(t, nexus.URL, hexis.URL)
+
+	hexis.SetFault(401)
+	reply := h.handleHexisAct(ctx, actDec("muzick indexer"))
+	if !denied(serviceHexis, reply) {
+		t.Fatalf("401 from hexis discovery: got %q, want the denied line naming Hexis", reply)
+	}
+	if !strings.Contains(reply, serviceHexis) {
+		t.Errorf("reply does not name Hexis: %q", reply)
+	}
+}
+
+// TestHexisDiscoveryOutageIsDownNotDenied — the other half of the same fork.
+// Without this the fix could pass by calling everything a refused credential.
+func TestHexisDiscoveryOutageIsDownNotDenied(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", muzickIndexer, "service"))
+	h := hexisGapHandler(t, nexus.URL, unreachableURL)
+
+	reply := h.handleHexisAct(ctx, actDec("muzick indexer"))
+	if !down(serviceHexis, reply) {
+		t.Fatalf("connection refused from hexis: got %q, want the outage line naming Hexis", reply)
+	}
+	if denied(serviceHexis, reply) {
+		t.Error("an outage must not be reported as a refused credential")
+	}
+}
+
+// TestHexisExecute401IsDeniedNotCommandFailure — the execute hop, which did not
+// consult ecosystemGap at all and named neither the service nor the cause.
+func TestHexisExecute401IsDeniedNotCommandFailure(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", muzickIndexer, "service"))
+	caps := fixtureHexisCapabilities(map[string]any{"id": "cap_restart", "name": "restart", "read_only": true})
+	hexis := newFakeHexis(t, caps, fixtureHexisExecuted("exec_1", "succeeded"))
+	h := hexisGapHandler(t, nexus.URL, hexis.URL)
+
+	// Discovery stays healthy; only the execute endpoint refuses. A blanket
+	// fault would never reach the site under test.
+	hexis.SetRouteFault("/api/v1/execute", 401)
+	reply := h.handleHexisAct(ctx, actDec("muzick indexer"))
+	if !denied(serviceHexis, reply) {
+		t.Fatalf("401 from hexis execute: got %q, want the denied line naming Hexis", reply)
+	}
+}
+
+// TestHexisExecuteOutageIsDown — same site, the other classification.
+//
+// Discovery and execution share one base URL, so the outage has to be scoped to
+// the execute endpoint rather than to the server: it answers capabilities
+// normally and drops the connection on execute, which is what the client sees
+// when the far side dies mid-call. That produces no HTTP status at all, which is
+// what Unreachable() means.
+func TestHexisExecuteOutageIsDown(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", muzickIndexer, "service"))
+	caps := fixtureHexisCapabilities(map[string]any{"id": "cap_restart", "name": "restart", "read_only": true})
+	hexis := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if r.URL.Path == "/api/v1/execute" {
+			conn, _, err := w.(http.Hijacker).Hijack()
+			if err != nil {
+				t.Errorf("hijack: %v", err)
+				return
+			}
+			conn.Close()
+			return
+		}
+		w.Header().Set("Content-Type", "application/json")
+		_, _ = w.Write([]byte(caps))
+	}))
+	t.Cleanup(hexis.Close)
+	h := hexisGapHandler(t, nexus.URL, hexis.URL)
+
+	reply := h.handleHexisAct(ctx, actDec("muzick indexer"))
+	if !down(serviceHexis, reply) {
+		t.Fatalf("dropped connection on hexis execute: got %q, want the outage line", reply)
+	}
+	if denied(serviceHexis, reply) {
+		t.Error("an outage must not be reported as a refused credential")
+	}
+}
+
+// TestHexisExecutionFailedStaysCommandFailure — the boundary of the fix. Hexis
+// answering 200 with a failed execution is the command failing, not Hexis
+// degrading, and it must keep the command-level line rather than accusing a
+// healthy service of being down.
+func TestHexisExecutionFailedStaysCommandFailure(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", muzickIndexer, "service"))
+	caps := fixtureHexisCapabilities(map[string]any{"id": "cap_restart", "name": "restart", "read_only": true})
+	hexis := newFakeHexis(t, caps, fixtureHexisExecutionFailed("exec_1", "unit refused to start"))
+	h := hexisGapHandler(t, nexus.URL, hexis.URL)
+
+	reply := h.handleHexisAct(ctx, actDec("muzick indexer"))
+	if down(serviceHexis, reply) || denied(serviceHexis, reply) {
+		t.Fatalf("a failed execution must not be reported as an ecosystem gap, got %q", reply)
+	}
+	if !phraser.IsA(phraser.ActFailEntity, map[string]string{"name": muzickIndexer}, reply) {
+		t.Fatalf("want the command-failure line, got %q", reply)
+	}
+}

cmd/mavend/ecosystem_harness_test.go

@@ -115,3 +115,83 @@ func TestFakeNexus_FaultInjectionThenRecovery(t *testing.T) {
 		t.Fatalf("expected success once nexus recovers, got %q", reply)
 	}
 }
+
+// TestPraxisEntityAttention_RemembersWhatItReadOut: the scoped digest is a list
+// she read out, so a positional follow-up must land on one of ITS items. It
+// surfaced them and remembered none, which left the previous digest live and
+// sent "отметь второй" at somebody else's item.
+func TestPraxisEntityAttention_RemembersWhatItReadOut(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", "Muzick indexer", "service"))
+	scoped := fixturePraxisAttentionScoped("ent_muzick",
+		map[string]any{"id": "item_scoped_1", "title": "indexer wedged"})
+	praxis := newFakePraxis(t, scoped)
+	h := ecoHandler(t, nexus, praxis, nil)
+
+	// A digest from an earlier turn, still the positional memory.
+	h.rememberSurfaced([]string{"item_stale"})
+
+	reply := h.handlePraxisAct(ctx, router.Decision{
+		Intent: router.IntentAct,
+		Slots:  router.Slots{Fn: "entity_attention", HasFn: true, Value: "muzick indexer"},
+	})
+	if !strings.Contains(reply, "indexer wedged") {
+		t.Fatalf("expected the scoped item to be read out, got %q", reply)
+	}
+
+	h.mu.Lock()
+	surfaced := append([]string(nil), h.surfacedItems...)
+	h.mu.Unlock()
+	if len(surfaced) != 1 || surfaced[0] != "item_scoped_1" {
+		t.Fatalf("scoped digest must replace the positional memory, got %v", surfaced)
+	}
+
+	// The follow-up resolves against what he just heard, not the stale list.
+	if reply := h.handlePraxisAct(ctx, praxisItemDec("resolve_item", "last")); reply == "" {
+		t.Fatal("positional follow-up should have been claimed by praxis")
+	}
+	var body string
+	for _, r := range praxis.Requests() {
+		if r.Method == "POST" && r.Path == "/api/v1/tools/resolve" {
+			body = string(r.Body)
+		}
+	}
+	if !strings.Contains(body, "item_scoped_1") {
+		t.Fatalf("resolve must transition the item she read out, posted %q", body)
+	}
+	if strings.Contains(body, "item_stale") {
+		t.Fatal("resolve transitioned an item from a previous digest")
+	}
+}
+
+// TestHexisConfirm_KeepsOneCorrelationIDPerAction: the confirm arrives on a
+// later turn with a context of its own. The contract mints one id per action,
+// so the execution it authorises must still be joinable to the resolve and the
+// discovery that proposed it — it recorded a fresh id and no causation at all.
+func TestHexisConfirm_KeepsOneCorrelationIDPerAction(t *testing.T) {
+	ctx := context.Background()
+	nexus := newFakeNexus(t, fixtureNexusResolved("ent_muzick", "Muzick indexer", "service"))
+	caps := fixtureHexisCapabilities(map[string]any{"id": "cap_restart", "name": "restart", "read_only": false})
+	hexis := newFakeHexis(t, caps, fixtureHexisExecuted("exec_1", "succeeded"))
+	h := ecoHandler(t, nexus, nil, hexis)
+
+	if reply := h.handleHexisAct(ctx, actDec("restart")); !strings.Contains(reply, "да") {
+		t.Fatalf("mutating capability must ask for confirmation, got %q", reply)
+	}
+	resolve := findTrace(t, h, "nexus", "resolve")
+	if resolve == nil || resolve.CorrelationID == "" {
+		t.Fatalf("expected a nexus resolve trace carrying a correlation id, got %+v", resolve)
+	}
+
+	if _, handled := h.resolveConfirm(ctx, "да"); !handled {
+		t.Fatal("confirm should have been claimed")
+	}
+	exec := findTrace(t, h, "hexis", "execute")
+	if exec == nil {
+		t.Fatal("expected a hexis execute trace")
+	}
+	if exec.CausationID != resolve.CorrelationID {
+		t.Fatalf("confirmed execution must cite the action that proposed it: causation %q, action %q",
+			exec.CausationID, resolve.CorrelationID)
+	}
+}

cmd/mavend/fact_subject_test.go

@@ -24,7 +24,7 @@ func TestApplyAction_FactCapture_QueuesEntityResolution(t *testing.T) {
 
 	emb := router.NewHashEmbedder(1024)
 	matcher := tool.NewMatcher(api)
-	rtr := buildRouter(emb, matcher, 0.55, nil)
+	rtr := buildRouter(emb, matcher, 0.55, nil, nil)
 
 	h := &reactiveHandler{
 		api:       api,

cmd/mavend/factenrichment.go

@@ -37,8 +37,8 @@ type factEnrichmentWorker struct {
 	nextTry map[int64]time.Time // fact id → earliest retry
 }
 
-// enrichmentScanLimit bounds how deep a single tick (or status report) walks
-// the pending queue looking for facts whose backoff has elapsed. The queue is
+// enrichmentScanLimit bounds how deep a single tick walks the pending queue
+// looking for facts whose backoff has elapsed. The queue is
 // ordered by id, so without a scan the oldest facts hold every batch slot
 // whether or not they are eligible, and one permanently failing fact stalls
 // every younger one behind it.
@@ -75,8 +75,8 @@ func newFactEnrichmentWorker(st *store.Store, eco *ecosystemWiring, interval tim
 // has been down all day must be visible as a backlog, not as facts that
 // silently never got tagged.
 //
-// All three numbers describe the same set of rows, the first
-// enrichmentScanLimit pending facts. Counting Pending over a thousand rows
+// All three numbers describe the same set of rows, whatever is still pending
+// out of the first enrichmentScanLimit facts. Counting Pending over a thousand rows
 // while counting InBackoff over the twenty that reached the head of a batch
 // described two different populations under one struct.
 type enrichmentStatus struct {
@@ -86,13 +86,22 @@ type enrichmentStatus struct {
 	Scanned     int // rows the other three counts were taken over
 }
 
+// status reads the queue and counts over it. For a caller with no batch in
+// hand — anything asking the worker how it is doing from outside the tick.
 func (w *factEnrichmentWorker) status(ctx context.Context) enrichmentStatus {
-	var st enrichmentStatus
 	pending, err := w.store.PendingFactResolutions(ctx, enrichmentScanLimit)
 	if err != nil {
 		log.Printf("factenrichment: status: %v", err)
-		return st
+		return enrichmentStatus{}
 	}
+	return w.statusOf(pending)
+}
+
+// statusOf counts over a batch the caller already has. The batch is the query
+// the tick already ran, so reporting the backlog costs no second read of the
+// scan limit — up to a thousand rows, on a database that serialises them.
+func (w *factEnrichmentWorker) statusOf(pending []store.Fact) enrichmentStatus {
+	var st enrichmentStatus
 	st.Pending = len(pending)
 	st.Scanned = len(pending)
 	w.mu.Lock()
@@ -144,17 +153,24 @@ func (w *factEnrichmentWorker) tick(ctx context.Context) {
 	}
 	w.forgetDeparted(pending)
 	skipped, failed, attempted := 0, 0, 0
+	// A resolved fact leaves the pending queue, so the batch in hand overstates
+	// the backlog by however many succeeded. Drop them here rather than
+	// re-reading the queue to find out.
+	remaining := make([]store.Fact, 0, len(pending))
 	for _, f := range pending {
 		if attempted >= w.batch {
-			break
+			remaining = append(remaining, f)
+			continue
 		}
 		if !w.due(f.ID) {
 			skipped++
+			remaining = append(remaining, f)
 			continue
 		}
 		attempted++
 		if !w.resolveOne(ctx, f) {
 			failed++
+			remaining = append(remaining, f)
 		}
 	}
 	if failed > 0 {
@@ -164,7 +180,7 @@ func (w *factEnrichmentWorker) tick(ctx context.Context) {
 	// Report the backlog every tick, not only when something failed: the
 	// stalled state worth seeing is the one where nothing failed because
 	// nothing was attempted.
-	if st := w.status(ctx); st.Pending > 0 {
+	if st := w.statusOf(remaining); st.Pending > 0 {
 		log.Printf("factenrichment: %d facts pending entity resolution, %d in backoff, worst attempt %d (scanned %d)",
 			st.Pending, st.InBackoff, st.MaxAttempts, st.Scanned)
 	}

cmd/mavend/factgate_test.go

@@ -20,7 +20,7 @@ func newFactGateHandler(t *testing.T, now time.Time) (*reactiveHandler, ipc.Core
 	h := &reactiveHandler{
 		api:       api,
 		recall:    recallWiring{embedder: emb, memStore: memory.NewInMemoryStore()},
-		router:    buildRouter(emb, tool.NewMatcher(api), 0.55, nil),
+		router:    buildRouter(emb, tool.NewMatcher(api), 0.55, nil, nil),
 		replier:   voice.NewStubReplier(),
 		now:       func() time.Time { return now },
 		dataStore: st,

cmd/mavend/main.go

@@ -68,6 +68,11 @@ import (
 	"github.com/kami/maven/internal/webauthn"
 )
 
+// stepUpTTL is how long one passkey assertion keeps the session stepped up.
+// Long enough for the unlock call that follows it, short enough that a walked
+// away laptop does not stay authorized.
+const stepUpTTL = 5 * time.Minute
+
 var errLocked = errors.New("mavend: daemon locked — complete passkey assertion first")
 
 // daemonLock tracks whether the daemon is in locked (pre-unlock) mode, and
@@ -247,44 +252,31 @@ func run(args []string) error {
 	// envelope per successful intake write.
 	coreFor := func() ipc.CoreAPI { return newIntakeAPI(ipc.NewStoreAPI(st), evBus, time.Now) }
 
+	// depsNow reads whatever the current path has wired. Both boot paths build
+	// the CoreAPI and start the workers from this one value, so neither can
+	// hold a field the other misses. See cmd/mavend/boot.go.
+	depsNow := func() bootDeps {
+		return bootDeps{
+			coreFor:    coreFor,
+			tl:         tl,
+			evBus:      evBus,
+			voiceW:     voiceW,
+			st:         st,
+			factWorker: factWorker,
+			evalWorker: evalWorker,
+			feedWkr:    feedWkr,
+			crawlWkr:   crawlWkr,
+		}
+	}
+
 	if !locked {
 		rules = wireRules(cfg)
-		gatherer = loop.NewGatherer(st, rules)
-		if cfg.QuietHours != nil {
-			gatherer.SetQuietHours(cfg.QuietHours.Start, cfg.QuietHours.End)
-		}
+		gatherer = wireGatherer(st, cfg, rules)
 
 		// phraser
-		phr = phraser.NewStub()
-		if cfg.Phraser != nil {
-			pc := phraser.Config{
-				ModelPath:    cfg.Phraser.ModelPath,
-				BinPath:      cfg.Phraser.BinPath,
-				Listen:       cfg.Phraser.Listen,
-				NGpuLayers:   cfg.Phraser.NGpuLayers,
-				NCtx:         cfg.Phraser.NCtx,
-				CacheRAMMiB:  cacheRAMMiB(cfg.Phraser.CacheRAMMiB),
-				Timeout:      time.Duration(cfg.Phraser.Timeout),
-				LLMNudges:    cfg.Phraser.LLMNudges,
-				ContextBlock: contextBlockFn(cfg, time.Now),
-			}
-			if pc.BinPath == "" {
-				pc.BinPath = "llama-server"
-			}
-			if pc.Listen == "" {
-				pc.Listen = "127.0.0.1:0"
-			}
-			if pc.NCtx <= 0 {
-				pc.NCtx = 2048
-			}
-			if pc.Timeout <= 0 {
-				pc.Timeout = 30 * time.Second
-			}
-			var err error
-			phr, err = phraser.NewLLMPhraser(ctx, pc)
-			if err != nil {
-				return fmt.Errorf("phraser: %w", err)
-			}
+		phr, err = wirePhraser(ctx, cfg)
+		if err != nil {
+			return fmt.Errorf("phraser: %w", err)
 		}
 
 		// ecosystem — nexus + hexis + praxis (all over HTTP; no direct DB access)
@@ -297,72 +289,19 @@ func run(args []string) error {
 		}
 
 		// delivery
-		var ntfy delivery.Sink
-		if cfg.Ntfy != nil {
-			s, err := ntfysink.New(*cfg.Ntfy)
-			if err != nil {
-				return fmt.Errorf("wire ntfy sink: %w", err)
-			}
-			ntfy = s
+		dispatcher, err = wireDispatcher(st, cfg, voiceW)
+		if err != nil {
+			return err
 		}
-		var telegram delivery.Sink
-		if cfg.Telegram != nil {
-			s, err := telegramsink.New(*cfg.Telegram)
-			if err != nil {
-				return fmt.Errorf("wire telegram sink: %w", err)
-			}
-			telegram = s
-		}
-		var voiceSink delivery.Sink
-		if voiceW != nil {
-			voiceSink = voiceW.voiceSink
-		}
-		// A crashed prior run may have left "pending" delivery attempts (send
-		// may have landed externally, then the process died before recording
-		// it) — reconcile them to "unknown" before the tick loop resumes
-		// sending, so nothing auto-resends into that ambiguity.
-		if _, err := st.ReconcileStaleDeliveryAttempts(context.Background(), time.Now()); err != nil {
-			log.Printf("delivery outbox reconcile: %v", err)
-		}
-		dispatcher = delivery.NewDispatcher(delivery.Config{
-			Ntfy:      ntfy,
-			Telegram:  telegram,
-			Voice:     voiceSink,
-			Ack:       st,
-			Nudges:    st,
-			Reminders: st,
-			Outbox:    st,
-		})
 
 		// tick loop
-		tickInterval := time.Duration(cfg.TickInterval)
-		repeatInterval := time.Duration(cfg.RepeatInterval)
-		autotuneInterval := time.Duration(cfg.AutotuneInterval)
-		tl = newTickLoop(st, gatherer, dispatcher, phr, rules, tickInterval, repeatInterval, autotuneInterval, cfg.Digest, routinesFromConfig(cfg.Routines), config.MorningRoutinesFromConfig(cfg.MorningRoutines), cfg.PatternProposals)
+		tl = wireTickLoop(st, gatherer, dispatcher, phr, rules, cfg)
 		factWorker = newFactEnrichmentWorker(st, eco, time.Duration(cfg.FactEnrichmentInterval))
 		evalWorker = newMemoryEvalWorker(st, phr, cfg)
 		feedWkr = newFeedWorker(coreFor(), embedderOf(voiceW), cfg)
 		crawlWkr = newCrawlWorker(newCrawler(cfg), coreFor(), embedderOf(voiceW), cfg)
 
-		coreAPI = &daemonAPI{
-			CoreAPI:          coreFor(),
-			getTrace:         tl.trace,
-			getMorningStatus: func(ctx context.Context) []ipc.MorningRoutineStatus { return tl.morningStatus(ctx, time.Now()) },
-			getDayPlan:       func(ctx context.Context) ipc.DayPlan { return tl.dayPlan(ctx, time.Now()) },
-			getEvents:        intakeEventsFn(evBus),
-			seedStore:        seedStoreIfAllowed(st),
-			nexus:            nexusOf(voiceW),
-		}
-		if voiceW != nil && voiceW.handler != nil {
-			api := coreAPI.(*daemonAPI)
-			api.chatFn = voiceW.handler.handleText
-			// And the reverse: the handler was wired with the bare store
-			// adapter, which cannot serve the day plan. See upgradeAPI.
-			voiceW.handler.upgradeAPI(api)
-		}
-		if voiceW != nil && voiceW.mcp != nil {
-			coreAPI.(*daemonAPI).getMCPServers = voiceW.mcp.status
-		}
+		coreAPI = newDaemonAPI(depsNow())
 	} else {
 		// locked mode: no real store yet, so there's no meaningful CoreAPI to
 		// serve. srv.Check below is the actual guard — every CoreAPI call is
@@ -379,7 +318,7 @@ func run(args []string) error {
 		return fmt.Errorf("ipc listen: %w", err)
 	}
 
-	passkeySess := webauthn.NewPasskeySession(5 * time.Minute)
+	passkeySess := webauthn.NewPasskeySession(stepUpTTL)
 
 	// Set Server.Check — the single authorization guard, run once by
 	// Server.dispatch before any CoreAPI method is called (see
@@ -423,6 +362,9 @@ func run(args []string) error {
 	if !locked {
 		wireMailIntake(srv, st, phr, cfg, evBus)
 		wireModelSwap(srv, phr, cfg)
+		// Inbound telegram (V-637). Dark unless the telegram block says intake,
+		// and it reads one chat.
+		wireTelegramIntake(ctx, &wg, coreAPI, cfg)
 		// Vision + the media blob store (Vikunja #252). Both stay dark without a
 		// media block; MethodDescribeImage answers ErrUnknownMethod then.
 		keeper := wireVision(ctx, &wg, srv, st, embedderOf(voiceW), cfg)
@@ -527,40 +469,11 @@ func run(args []string) error {
 
 			// Wire everything.
 			rules = wireRules(cfg)
-			gatherer = loop.NewGatherer(st, rules)
-			if cfg.QuietHours != nil {
-				gatherer.SetQuietHours(cfg.QuietHours.Start, cfg.QuietHours.End)
-			}
+			gatherer = wireGatherer(st, cfg, rules)
 
-			phr = phraser.NewStub()
-			if cfg.Phraser != nil {
-				pc := phraser.Config{
-					ModelPath:    cfg.Phraser.ModelPath,
-					BinPath:      cfg.Phraser.BinPath,
-					Listen:       cfg.Phraser.Listen,
-					NGpuLayers:   cfg.Phraser.NGpuLayers,
-					NCtx:         cfg.Phraser.NCtx,
-					CacheRAMMiB:  cacheRAMMiB(cfg.Phraser.CacheRAMMiB),
-					Timeout:      time.Duration(cfg.Phraser.Timeout),
-					LLMNudges:    cfg.Phraser.LLMNudges,
-					ContextBlock: contextBlockFn(cfg, time.Now),
-				}
-				if pc.BinPath == "" {
-					pc.BinPath = "llama-server"
-				}
-				if pc.Listen == "" {
-					pc.Listen = "127.0.0.1:0"
-				}
-				if pc.NCtx <= 0 {
-					pc.NCtx = 2048
-				}
-				if pc.Timeout <= 0 {
-					pc.Timeout = 30 * time.Second
-				}
-				phr, err = phraser.NewLLMPhraser(ctx, pc)
-				if err != nil {
-					return fmt.Errorf("phraser: %w", err)
-				}
+			phr, err = wirePhraser(ctx, cfg)
+			if err != nil {
+				return fmt.Errorf("phraser: %w", err)
 			}
 
 			eco = wireEcosystem(cfg)
@@ -570,65 +483,26 @@ func run(args []string) error {
 				return fmt.Errorf("wire voice: %w", err)
 			}
 
-			var ntfy delivery.Sink
-			if cfg.Ntfy != nil {
-				s, err := ntfysink.New(*cfg.Ntfy)
-				if err != nil {
-					return fmt.Errorf("wire ntfy sink: %w", err)
-				}
-				ntfy = s
+			dispatcher, err = wireDispatcher(st, cfg, voiceW)
+			if err != nil {
+				return err
 			}
-			var telegram delivery.Sink
-			if cfg.Telegram != nil {
-				s, err := telegramsink.New(*cfg.Telegram)
-				if err != nil {
-					return fmt.Errorf("wire telegram sink: %w", err)
-				}
-				telegram = s
-			}
-			var voiceSink delivery.Sink
-			if voiceW != nil {
-				voiceSink = voiceW.voiceSink
-			}
-			if _, err := st.ReconcileStaleDeliveryAttempts(context.Background(), time.Now()); err != nil {
-				log.Printf("delivery outbox reconcile: %v", err)
-			}
-			dispatcher = delivery.NewDispatcher(delivery.Config{
-				Ntfy:      ntfy,
-				Telegram:  telegram,
-				Voice:     voiceSink,
-				Ack:       st,
-				Nudges:    st,
-				Reminders: st,
-				Outbox:    st,
-			})
 
-			tickInterval := time.Duration(cfg.TickInterval)
-			repeatInterval := time.Duration(cfg.RepeatInterval)
-			autotuneInterval := time.Duration(cfg.AutotuneInterval)
-			tl = newTickLoop(st, gatherer, dispatcher, phr, rules, tickInterval, repeatInterval, autotuneInterval, cfg.Digest, routinesFromConfig(cfg.Routines), config.MorningRoutinesFromConfig(cfg.MorningRoutines), cfg.PatternProposals)
+			tl = wireTickLoop(st, gatherer, dispatcher, phr, rules, cfg)
 			factWorker = newFactEnrichmentWorker(st, eco, time.Duration(cfg.FactEnrichmentInterval))
 			evalWorker = newMemoryEvalWorker(st, phr, cfg)
 			feedWkr = newFeedWorker(coreFor(), embedderOf(voiceW), cfg)
 			crawlWkr = newCrawlWorker(newCrawler(cfg), coreFor(), embedderOf(voiceW), cfg)
 
 			// Swap the CoreAPI from the locked placeholder to the real store adapter.
-			newAPI := &daemonAPI{
-				CoreAPI:          coreFor(),
-				getTrace:         tl.trace,
-				getMorningStatus: func(ctx context.Context) []ipc.MorningRoutineStatus { return tl.morningStatus(ctx, time.Now()) },
-				getDayPlan:       func(ctx context.Context) ipc.DayPlan { return tl.dayPlan(ctx, time.Now()) },
-				getEvents:        intakeEventsFn(evBus),
-				seedStore:        seedStoreIfAllowed(st),
-			}
-			if voiceW != nil && voiceW.handler != nil {
-				newAPI.chatFn = voiceW.handler.handleText
-				voiceW.handler.upgradeAPI(newAPI)
-			}
+			newAPI := newDaemonAPI(depsNow())
 			srv.SetAPI(newAPI)
 			srv.Check = (&auth.Gate{Enrollment: auth.NewFloorEnrollment(), Session: passkeySess}).Check
 			wireMailIntake(srv, st, phr, cfg, evBus)
 			wireModelSwap(srv, phr, cfg)
+			// Same on the unlock path, with the API that has just replaced the
+			// locked placeholder (V-637).
+			wireTelegramIntake(ctx, &wg, newAPI, cfg)
 			keeper := wireVision(ctx, &wg, srv, st, embedderOf(voiceW), cfg)
 			wireCapture(ctx, &wg, srv, keeper, st, voiceW, phr, cfg)
 			// Voice identification (Vikunja #255). Enrolment plumbing only until a
@@ -636,59 +510,10 @@ func run(args []string) error {
 			// block, so no wire path takes a voiceprint on a default box.
 			wireSpeaker(srv, st, cfg)
 
-			// Start voice server.
-			if voiceW != nil {
-				var wg sync.WaitGroup
-				wg.Add(1)
-				go func() {
-					defer wg.Done()
-					if err := voiceW.server.Serve(); err != nil && !errors.Is(err, net.ErrClosed) {
-						log.Printf("voice serve: %v", err)
-					}
-				}()
-				log.Printf("mavend: voice listening on %s", voiceW.server.Addr())
-			}
-
-			// Start tick loop.
-			go func() {
-				tl.run(ctx)
-			}()
-
-			// Start fact-entity enrichment worker.
-			go func() {
-				factWorker.run(ctx)
-			}()
-
-			// Start background memory evaluation (nil unless configured).
-			if evalWorker != nil {
-				go func() {
-					evalWorker.run(ctx)
-				}()
-			}
-
-			// Start feed reading (nil unless configured).
-			if feedWkr != nil {
-				go func() {
-					feedWkr.run(ctx)
-				}()
-			}
-
-			// Start the watched-page crawls (nil unless configured).
-			if crawlWkr != nil {
-				go func() {
-					crawlWkr.run(ctx)
-				}()
-			}
-
-			// Keep MCP connections alive (nil unless configured).
-			if voiceW != nil && voiceW.mcp != nil {
-				go voiceW.mcp.run(ctx)
-			}
-
-			// Re-enumerate the house for new devices (nil unless configured).
-			if voiceW != nil && voiceW.home != nil {
-				go voiceW.home.run(ctx)
-			}
+			// The voice server and every background worker, on the outer wg
+			// so shutdown waits for them. This used to be nine bare
+			// `go func()` calls and a shadowed WaitGroup (V-639).
+			startBackground(ctx, &wg, depsNow())
 
 			dl.unlock(st)
 			log.Printf("mavend: unlocked via passkey assertion")
@@ -696,72 +521,15 @@ func run(args []string) error {
 		}
 	}
 
-	wg.Add(1)
-	go func() {
-		defer wg.Done()
+	goWorker(&wg, func() {
 		if err := srv.Serve(); err != nil && !errors.Is(err, net.ErrClosed) {
 			log.Printf("ipc serve: %v", err)
 		}
-	}()
+	})
 	log.Printf("mavend: ipc listening on %s", srv.Path())
 
-	if !locked && voiceW != nil {
-		wg.Add(1)
-		go func() {
-			defer wg.Done()
-			if err := voiceW.server.Serve(); err != nil && !errors.Is(err, net.ErrClosed) {
-				log.Printf("voice serve: %v", err)
-			}
-		}()
-		log.Printf("mavend: voice listening on %s", voiceW.server.Addr())
-	}
-
 	if !locked {
-		wg.Add(1)
-		go func() {
-			defer wg.Done()
-			tl.run(ctx)
-		}()
-		wg.Add(1)
-		go func() {
-			defer wg.Done()
-			factWorker.run(ctx)
-		}()
-		if evalWorker != nil {
-			wg.Add(1)
-			go func() {
-				defer wg.Done()
-				evalWorker.run(ctx)
-			}()
-		}
-		if feedWkr != nil {
-			wg.Add(1)
-			go func() {
-				defer wg.Done()
-				feedWkr.run(ctx)
-			}()
-		}
-		if crawlWkr != nil {
-			wg.Add(1)
-			go func() {
-				defer wg.Done()
-				crawlWkr.run(ctx)
-			}()
-		}
-		if voiceW != nil && voiceW.mcp != nil {
-			wg.Add(1)
-			go func() {
-				defer wg.Done()
-				voiceW.mcp.run(ctx)
-			}()
-		}
-		if voiceW != nil && voiceW.home != nil {
-			wg.Add(1)
-			go func() {
-				defer wg.Done()
-				voiceW.home.run(ctx)
-			}()
-		}
+		startBackground(ctx, &wg, depsNow())
 	}
 
 	<-ctx.Done()
@@ -848,6 +616,125 @@ func waitWorkers(wg *sync.WaitGroup, d time.Duration) bool {
 	}
 }
 
+// Phraser defaults, applied when the config block leaves a field unset. They
+// are the daemon's, not the library's: phraser.Config carries no defaults of
+// its own, so an empty field here would reach llama-server as an empty flag.
+const (
+	defaultLlamaBin       = "llama-server"
+	defaultPhraserListen  = "127.0.0.1:0"
+	defaultPhraserNCtx    = 2048
+	defaultPhraserTimeout = 30 * time.Second
+)
+
+// wirePhraser builds the phrasing seam. No phraser block means the
+// deterministic stub, which is the floor and not an error: the daemon answers
+// without a model, in fixed words.
+func wirePhraser(ctx context.Context, cfg *config.Config) (phraser.Phraser, error) {
+	if cfg.Phraser == nil {
+		return phraser.NewStub(), nil
+	}
+	pc := phraser.Config{
+		ModelPath:    cfg.Phraser.ModelPath,
+		BinPath:      cfg.Phraser.BinPath,
+		Listen:       cfg.Phraser.Listen,
+		NGpuLayers:   cfg.Phraser.NGpuLayers,
+		NCtx:         cfg.Phraser.NCtx,
+		CacheRAMMiB:  cacheRAMMiB(cfg.Phraser.CacheRAMMiB),
+		Timeout:      time.Duration(cfg.Phraser.Timeout),
+		LLMNudges:    cfg.Phraser.LLMNudges,
+		ContextBlock: contextBlockFn(cfg, time.Now),
+	}
+	if pc.BinPath == "" {
+		pc.BinPath = defaultLlamaBin
+	}
+	if pc.Listen == "" {
+		pc.Listen = defaultPhraserListen
+	}
+	if pc.NCtx <= 0 {
+		pc.NCtx = defaultPhraserNCtx
+	}
+	if pc.Timeout <= 0 {
+		pc.Timeout = defaultPhraserTimeout
+	}
+	return phraser.NewLLMPhraser(ctx, pc)
+}
+
+// wireGatherer builds the nudge gatherer over the given rule set and applies
+// the configured quiet hours.
+func wireGatherer(st *store.Store, cfg *config.Config, rules []loop.Rule) *loop.Gatherer {
+	g := loop.NewGatherer(st, rules)
+	if cfg.QuietHours != nil {
+		g.SetQuietHours(cfg.QuietHours.Start, cfg.QuietHours.End)
+	}
+	return g
+}
+
+// wireDispatcher builds the delivery fan-out. Each sink stays nil unless its
+// config block is present, and a sink that fails to build fails the boot
+// rather than going quiet.
+//
+// A crashed prior run may have left "pending" delivery attempts (send may have
+// landed externally, then the process died before recording it). They are
+// reconciled to "unknown" here, before the tick loop resumes sending, so
+// nothing auto-resends into that ambiguity.
+func wireDispatcher(st *store.Store, cfg *config.Config, voiceW *voiceWiring) (*delivery.Dispatcher, error) {
+	var ntfy delivery.Sink
+	if cfg.Ntfy != nil {
+		s, err := ntfysink.New(*cfg.Ntfy)
+		if err != nil {
+			return nil, fmt.Errorf("wire ntfy sink: %w", err)
+		}
+		ntfy = s
+	}
+	var telegram delivery.Sink
+	if cfg.Telegram != nil {
+		s, err := telegramsink.New(*cfg.Telegram)
+		if err != nil {
+			return nil, fmt.Errorf("wire telegram sink: %w", err)
+		}
+		telegram = s
+	}
+	var voiceSink delivery.Sink
+	if voiceW != nil {
+		voiceSink = voiceW.voiceSink
+	}
+	if _, err := st.ReconcileStaleDeliveryAttempts(context.Background(), time.Now()); err != nil {
+		log.Printf("delivery outbox reconcile: %v", err)
+	}
+	return delivery.NewDispatcher(delivery.Config{
+		Ntfy:      ntfy,
+		Telegram:  telegram,
+		Voice:     voiceSink,
+		Ack:       st,
+		Nudges:    st,
+		Reminders: st,
+		Outbox:    st,
+	}), nil
+}
+
+// wireTickLoop reads the loop's three intervals and its schedules out of the
+// config, so the two boot paths cannot disagree about them.
+func wireTickLoop(st *store.Store, gatherer *loop.Gatherer, dispatcher *delivery.Dispatcher, phr phraser.Phraser, rules []loop.Rule, cfg *config.Config) *tickLoop {
+	return newTickLoop(st, gatherer, dispatcher, phr, rules,
+		time.Duration(cfg.TickInterval),
+		time.Duration(cfg.RepeatInterval),
+		time.Duration(cfg.AutotuneInterval),
+		cfg.Digest,
+		routinesFromConfig(cfg.Routines),
+		config.MorningRoutinesFromConfig(cfg.MorningRoutines),
+		cfg.PatternProposals)
+}
+
+// goWorker starts run on its own goroutine and registers it with wg, so
+// shutdown can wait for it inside workerGrace.
+func goWorker(wg *sync.WaitGroup, run func()) {
+	wg.Add(1)
+	go func() {
+		defer wg.Done()
+		run()
+	}()
+}
+
 // wireRules builds the nudge rule set, minus anything config turned off. The
 // drop is logged because a rule vanishing silently is indistinguishable from a
 // rule that is broken, and the next person to wonder why she stopped nudging

cmd/mavend/notefragment.go

@@ -0,0 +1,58 @@
+package main
+
+import (
+	"strings"
+
+	"github.com/kami/maven/internal/lexicon"
+	"github.com/kami/maven/internal/morph"
+)
+
+// correctionFragment reports that an utterance replaces a referent and states
+// nothing of its own: "нет, не маме, а папе" (V-576).
+//
+// Measured on the box 2026-08-06, that fragment routed to note and was filed.
+// It is not a repair either, because it names no intent, so parseRepair
+// declines it and repair.go is the wrong place to catch it. This is the note
+// path saying it has nothing to store.
+//
+// Three offline tests, all of them narrow on purpose. The sentence opens with a
+// refusal word from the lexicon, it carries the contrastive "а" that names the
+// replacement, and no token in it is a verb form. The verb test is what keeps
+// the rule off real notes: "нет, я не поеду, а останусь" says something, and a
+// Russian verb carries its own subject and tense.
+func correctionFragment(utterance string) bool {
+	toks := repairTokens(strings.ToLower(strings.TrimSpace(utterance)))
+	if len(toks) < 3 {
+		return false
+	}
+	if !refusalWord(toks[0]) {
+		return false
+	}
+	var negated, contrasted bool
+	for _, tok := range toks[1:] {
+		switch tok {
+		case "не", "not":
+			negated = true
+		case "а", "but":
+			contrasted = true
+		}
+		if morph.IsVerbForm(tok) {
+			return false
+		}
+	}
+	return negated && contrasted
+}
+
+// refusalWord reports that a token is a one-word refusal. The lexicon set holds
+// phrases too ("не надо"), and those are not what opens a correction.
+func refusalWord(tok string) bool {
+	for _, w := range lexicon.ConfirmNo() {
+		if strings.ContainsRune(w, ' ') {
+			continue
+		}
+		if w == tok {
+			return true
+		}
+	}
+	return false
+}

cmd/mavend/notefragment_test.go

@@ -0,0 +1,117 @@
+package main
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/memory"
+	"github.com/kami/maven/internal/router"
+	"github.com/kami/maven/internal/store"
+	"github.com/kami/maven/internal/tool"
+	"github.com/kami/maven/internal/voice"
+)
+
+func TestCorrectionFragment(t *testing.T) {
+	cases := []struct {
+		utterance string
+		want      bool
+	}{
+		{"нет, не маме, а папе", true},
+		{"Нет, не маме — а папе", true},
+		{"no, not mom, but dad", true},
+		// States something of its own, so it is his to keep.
+		{"нет, я не поеду, а останусь дома", false},
+		{"нет", false},
+		{"не маме, а папе", false},  // no refusal word opening it
+		{"нет, маме и папе", false}, // nothing negated
+		{"нет, не маме", false},     // nothing put in its place
+		{"запомни что кофе закончился", false},
+	}
+	for _, c := range cases {
+		if got := correctionFragment(c.utterance); got != c.want {
+			t.Errorf("correctionFragment(%q) = %v, want %v", c.utterance, got, c.want)
+		}
+	}
+}
+
+func newNoteHandler(t *testing.T) (*reactiveHandler, *store.Store) {
+	t.Helper()
+	st := newTestStore(t)
+	api := ipc.NewStoreAPI(st)
+	now := time.Now()
+	emb := router.NewHashEmbedder(1024)
+	h := &reactiveHandler{
+		api:       api,
+		recall:    recallWiring{embedder: emb, memStore: memory.NewInMemoryStore()},
+		router:    buildRouter(emb, tool.NewMatcher(api), 0.55, nil, nil),
+		replier:   voice.NewStubReplier(),
+		now:       func() time.Time { return now },
+		dataStore: st,
+	}
+	return h, st
+}
+
+// TestNoteBodyIsTheUtterance — the stored body comes from the utterance, never
+// from Slots.Text, which the LLM router is free to write anything into (V-576).
+func TestNoteBodyIsTheUtterance(t *testing.T) {
+	ctx := context.Background()
+	h, st := newNoteHandler(t)
+
+	dec := router.Decision{
+		Intent:    router.IntentNote,
+		Utterance: "купил хлеб и молоко",
+		Slots:     router.Slots{Text: "ты поедешь на дачу"},
+	}
+	if reply := h.applyAction(ctx, dec); reply != "" {
+		t.Fatalf("applyAction = %q, want empty", reply)
+	}
+	notes, err := st.RecentNotes(ctx, 10)
+	if err != nil {
+		t.Fatalf("RecentNotes: %v", err)
+	}
+	if len(notes) != 1 || notes[0].Text != dec.Utterance {
+		t.Fatalf("stored note = %+v, want body %q", notes, dec.Utterance)
+	}
+}
+
+// TestNoteBodyIsStable — the same utterance twice stores the same text.
+func TestNoteBodyIsStable(t *testing.T) {
+	ctx := context.Background()
+	h, st := newNoteHandler(t)
+
+	dec := router.Decision{Intent: router.IntentNote, Utterance: "кофе закончился"}
+	h.applyAction(ctx, dec)
+	h.applyAction(ctx, dec)
+
+	notes, err := st.RecentNotes(ctx, 10)
+	if err != nil {
+		t.Fatalf("RecentNotes: %v", err)
+	}
+	if len(notes) != 2 {
+		t.Fatalf("notes = %d, want 2", len(notes))
+	}
+	if notes[0].Text != notes[1].Text || notes[0].Text != dec.Utterance {
+		t.Fatalf("bodies differ: %q vs %q", notes[0].Text, notes[1].Text)
+	}
+}
+
+// TestCorrectionFragmentWritesNoNote — a correction with nothing behind it is
+// not a note, and she says so instead of filing it (V-576).
+func TestCorrectionFragmentWritesNoNote(t *testing.T) {
+	ctx := context.Background()
+	h, st := newNoteHandler(t)
+
+	dec := router.Decision{Intent: router.IntentNote, Utterance: "нет, не маме, а папе"}
+	if reply := h.applyAction(ctx, dec); reply != nothingToCorrectReply {
+		t.Fatalf("reply = %q, want %q", reply, nothingToCorrectReply)
+	}
+	notes, err := st.RecentNotes(ctx, 10)
+	if err != nil {
+		t.Fatalf("RecentNotes: %v", err)
+	}
+	if len(notes) != 0 {
+		t.Fatalf("notes = %+v, want none", notes)
+	}
+}

cmd/mavend/praxis_gap_test.go

@@ -0,0 +1,54 @@
+package main
+
+import (
+	"context"
+	"strings"
+	"testing"
+
+	"github.com/kami/maven/internal/phraser"
+)
+
+// A Praxis lifecycle failure used to return a hardcoded constant that named the
+// verb and never the service, so an outage, a refused token and a contract
+// mismatch all produced the identical sentence (Vikunja #588). The helpers and
+// the Hexis half of the same defect are in ecosystem_gap_test.go.
+
+func TestPraxisLifecycle401NamesPraxis(t *testing.T) {
+	ctx := context.Background()
+	praxis := newFakePraxis(t, fixturePraxisAttentionItems())
+	h := newPraxisTestHandler(t, praxis)
+
+	praxis.SetFault(401)
+	reply := h.handlePraxisAct(ctx, praxisItemDec("resolve_item", "item_1"))
+	if !strings.Contains(reply, servicePraxis) {
+		t.Fatalf("praxis failure does not name Praxis: %q", reply)
+	}
+	if !strings.Contains(reply, phraser.A(phraser.EcoDenied, serviceVars(servicePraxis))) {
+		t.Fatalf("401 from praxis: got %q, want the denied line", reply)
+	}
+	// The verb that did not happen is still said: the trace is the only other
+	// place it exists and he is not reading the trace.
+	if !strings.Contains(reply, "не получилось отметить сделанным.") {
+		t.Errorf("reply dropped the operation that failed: %q", reply)
+	}
+}
+
+// TestPraxisLifecycleOutageDiffersFrom401 — the identity that was the bug.
+func TestPraxisLifecycleOutageDiffersFrom401(t *testing.T) {
+	ctx := context.Background()
+	praxis := newFakePraxis(t, fixturePraxisAttentionItems())
+	h := newPraxisTestHandler(t, praxis)
+
+	praxis.SetFault(401)
+	refused := h.handlePraxisAct(ctx, praxisItemDec("acknowledge_item", "item_1"))
+
+	h.ecosystem = &ecosystemWiring{praxis: newPraxisClient(unreachableURL)}
+	outage := h.handlePraxisAct(ctx, praxisItemDec("acknowledge_item", "item_1"))
+
+	if refused == outage {
+		t.Fatalf("a refused token and an outage still say the same thing: %q", refused)
+	}
+	if !strings.Contains(outage, phraser.A(phraser.EcoDown, serviceVars(servicePraxis))) {
+		t.Fatalf("praxis outage: got %q, want the outage line naming Praxis", outage)
+	}
+}

cmd/mavend/query_recall_test.go

@@ -2,8 +2,11 @@ package main
 
 import (
 	"context"
+	"errors"
 	"fmt"
 	"math"
+	"net/http"
+	"strings"
 	"testing"
 	"time"
 
@@ -31,6 +34,57 @@ func (f *fixedEmbedder) Embed(_ context.Context, text string) ([]float32, error)
 	return v, nil
 }
 
+// brokenEmbedder fails every call, which is what an ONNX session error looks
+// like from the query chain's side.
+type brokenEmbedder struct{}
+
+func (brokenEmbedder) Dim() int     { return 4 }
+func (brokenEmbedder) Close() error { return nil }
+func (brokenEmbedder) Embed(context.Context, string) ([]float32, error) {
+	return nil, errors.New("onnx: session failed")
+}
+
+// TestQueryEmbedFailureDoesNotStopTheChain — V-568. The embed source used to
+// claim the turn on an embedder error, so one failing EmbedQuery answered every
+// question below it with "не смогла ответить", including the ones the search
+// answers without an embedder at all. A source that could not look must pass.
+func TestQueryEmbedFailureDoesNotStopTheChain(t *testing.T) {
+	const q = "почему небо голубое"
+
+	h, _ := searchHandler(t, searchBody, http.StatusOK)
+	h.api = ipc.NewStoreAPI(newTestStore(t))
+	h.recall = recallWiring{embedder: brokenEmbedder{}, minScore: 0.55, minMargin: 0.008}
+	h.now = time.Now
+
+	// The embed source itself passes rather than claiming.
+	turn := &queryTurn{dec: router.Decision{Intent: router.IntentQuery, Utterance: q}}
+	if reply, ok := h.queryEmbed(context.Background(), turn); ok {
+		t.Fatalf("queryEmbed claimed the turn on an embedder error: %q", reply)
+	}
+
+	// And the whole chain still reaches the search below it.
+	reply := askQuery(t, h, q)
+	if !strings.Contains(reply, "рэлеевского рассеяния") {
+		t.Fatalf("reply = %q, want the search answer", reply)
+	}
+}
+
+// The recall sources read the empty vector the failed embed left behind, and
+// neither of them may turn that into an answer: no vector means they could not
+// look, which is not the same as looking and finding nothing.
+func TestQueryRecallPassesWithoutAVector(t *testing.T) {
+	h, _ := buildRecallHandler(t, "где молоко", []recallCase{
+		{text: "молоко стоит в холодильнике", score: 0.90, kind: "note"},
+	})
+	turn := &queryTurn{dec: router.Decision{Intent: router.IntentQuery, Utterance: "где молоко"}}
+	if reply, ok := h.queryMemory(context.Background(), turn); ok {
+		t.Errorf("queryMemory claimed with no vector: %q", reply)
+	}
+	if reply, ok := h.queryNotes(context.Background(), turn); ok {
+		t.Errorf("queryNotes claimed with no vector: %q", reply)
+	}
+}
+
 // scoreVec builds a unit vector whose cosine against the query vector
 // (1,0,0,0) is exactly score.
 func scoreVec(score float64) []float32 {

cmd/mavend/querywalk_test.go

@@ -0,0 +1,137 @@
+package main
+
+import (
+	"testing"
+
+	"github.com/kami/maven/internal/router"
+)
+
+// The floor, and it is the reason a destination is safe to add at all: a box
+// whose model is down names nothing, and naming nothing has to walk the chain
+// the way it walked before the field existed.
+func TestNoDestinationWalksTheWholeChain(t *testing.T) {
+	walk, skipped := queryWalk(router.SourceUnknown, false)
+	if len(skipped) != 0 {
+		t.Errorf("skipped %d sources with no destination named, want none", len(skipped))
+	}
+	if len(walk) != len(querySources) {
+		t.Fatalf("walk has %d sources, want the whole table of %d", len(walk), len(querySources))
+	}
+	for i := range walk {
+		if walk[i].name != querySources[i].name {
+			t.Fatalf("position %d is %q, want %q", i, walk[i].name, querySources[i].name)
+		}
+	}
+}
+
+// The 2026-08-07 defects, one per line. Each is a source that decides by seed
+// similarity claiming a turn that was never its own, and then answering it
+// because it has no lookup that could come back empty.
+func TestANamedDestinationSilencesTheOtherGuessers(t *testing.T) {
+	cases := []struct {
+		dest      router.Source
+		utterance string
+		silenced  string
+		anchored  bool // a stage 0 grammar named the destination
+	}{
+		{router.SourceWorld, "что такое TCP?", "weather", true},
+		{router.SourceWorld, "сколько будет 17 на 23?", "weather", true},
+		{router.SourceWorld, "кто такой Линус Торвальдс?", "personal", true},
+		{router.SourceRecall, "какой у меня любимый язык?", "feeds", false},
+		{router.SourceCalendar, "что в календаре на завтра?", "weather", true},
+	}
+	for _, c := range cases {
+		walk, skipped := queryWalk(c.dest, c.anchored)
+		if inWalk(walk, c.silenced) {
+			t.Errorf("%q named %q: %q is still asked", c.utterance, c.dest, c.silenced)
+		}
+		if !inWalk(skipped, c.silenced) {
+			t.Errorf("%q named %q: %q is missing from the record of who was skipped",
+				c.utterance, c.dest, c.silenced)
+		}
+	}
+}
+
+// Naming the world must not send the turn outside. His notes, his facts and the
+// boundary in front of them are the invariant CLAUDE.md states as "the owner's
+// data first, then the world", and a destination a model wrote must not be able
+// to reverse it.
+func TestNamingTheWorldStillReadsHisDataFirst(t *testing.T) {
+	walk, _ := queryWalk(router.SourceWorld, true)
+	for _, look := range []string{"fact-by-key", "embed", "memory", "notes"} {
+		if !inWalk(walk, look) {
+			t.Errorf("%q was dropped; only the sources that guess may be dropped", look)
+		}
+	}
+	if posOf(walk, "notes") > posOf(walk, "search") {
+		t.Error("search is asked before his notes are")
+	}
+	if posOf(walk, "search") < 0 {
+		t.Fatal("search is not in the walk at all")
+	}
+}
+
+// The boundary belongs to his data, so naming recall keeps it. That is what
+// makes "какой у меня любимый язык?" answer "не нашла у тебя такой записи"
+// rather than reaching SearXNG once nothing local had it.
+func TestNamingRecallKeepsTheBoundary(t *testing.T) {
+	walk, _ := queryWalk(router.SourceRecall, true)
+	if !inWalk(walk, "personal") {
+		t.Fatal("the personal boundary was skipped on a turn named for his own data")
+	}
+	if posOf(walk, "personal") > posOf(walk, "search") {
+		t.Error("the boundary no longer sits in front of the world")
+	}
+}
+
+// The owner's call of 2026-08-09 (V-666): only a stage 0 grammar may take the
+// personal boundary off a turn. The routing heads and the resident model both
+// name a destination by inference, and an inferred SourceWorld would send a
+// question about him upstream. Every other guesser still goes.
+func TestOnlyAGrammarMayDropTheBoundary(t *testing.T) {
+	walk, skipped := queryWalk(router.SourceWorld, false)
+	if !inWalk(walk, "personal") {
+		t.Error("an inferred destination took the boundary off the turn")
+	}
+	if !inWalk(skipped, "weather") {
+		t.Error("weather is still asked; the rule covers the boundary alone")
+	}
+	if posOf(walk, "personal") > posOf(walk, "search") {
+		t.Error("the boundary no longer sits in front of the world")
+	}
+	if anchored, _ := queryWalk(router.SourceWorld, true); inWalk(anchored, "personal") {
+		t.Error(`a grammar named the world and the boundary stayed: ` +
+			`"кто такой Линус Торвальдс?" is answered "не нашла у тебя такой записи" again`)
+	}
+}
+
+// Whatever the destination, the walk is a subsequence of the table. Every
+// comment on that table argues an order between two sources, and none of those
+// reasons is about this field.
+func TestTheWalkNeverReordersTheTable(t *testing.T) {
+	for _, dest := range append([]router.Source{router.SourceUnknown}, router.Sources...) {
+		walk, skipped := queryWalk(dest, true)
+		if len(walk)+len(skipped) != len(querySources) {
+			t.Errorf("%q: %d walked + %d skipped, want %d", dest, len(walk), len(skipped), len(querySources))
+		}
+		last := -1
+		for _, s := range walk {
+			at := posOf(querySources, s.name)
+			if at <= last {
+				t.Errorf("%q: %q is out of table order", dest, s.name)
+			}
+			last = at
+		}
+	}
+}
+
+func inWalk(list []querySource, name string) bool { return posOf(list, name) >= 0 }
+
+func posOf(list []querySource, name string) int {
+	for i, s := range list {
+		if s.name == name {
+			return i
+		}
+	}
+	return -1
+}

cmd/mavend/reactive_notes_test.go

@@ -22,7 +22,7 @@ func TestReactiveNotesReminders(t *testing.T) {
 
 	emb := router.NewHashEmbedder(1024)
 	matcher := tool.NewMatcher(api)
-	rtr := buildRouter(emb, matcher, 0.55, nil)
+	rtr := buildRouter(emb, matcher, 0.55, nil, nil)
 
 	h := &reactiveHandler{
 		api:       api,
@@ -104,7 +104,7 @@ func TestSpokenTaskCaptureFilesATask(t *testing.T) {
 	h := &reactiveHandler{
 		api:       api,
 		recall:    recallWiring{embedder: emb, memStore: memory.NewInMemoryStore()},
-		router:    buildRouter(emb, matcher, 0.55, nil),
+		router:    buildRouter(emb, matcher, 0.55, nil, nil),
 		replier:   voice.NewStubReplier(),
 		now:       func() time.Time { return now },
 		dataStore: st,

cmd/mavend/reminderwhen.go

@@ -0,0 +1,209 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"strings"
+	"time"
+
+	"github.com/kami/maven/internal/dialogue"
+	"github.com/kami/maven/internal/router"
+)
+
+// A reminder commits only when three things are answered: what to say, what
+// time to say it, and what day (owner's rule, 2026-08-06, V-579). Anything
+// missing is asked about, and nothing missing is filled from the clock.
+//
+// "напомни завтра в 3 заказать цветы" has the what and the day and an hour that
+// could be either half of the day, so she asks which 3. "напомни в 9 вечера
+// разгрузить стиралку" has the what and an unambiguous hour and no day, so she
+// asks which day. Today being a valid reading is not the same as him saying it.
+//
+// Two things are already whole and are not asked about. A time that admits one
+// reading is not queried for its half of the day, so "завтра в 15:00" commits.
+// And an interval is an instant, so "через час" carries all three by itself.
+type whenGap string
+
+const (
+	whenComplete      whenGap = ""
+	whenNoHour        whenGap = "hour"
+	whenAmbiguousHour whenGap = "part_of_day"
+	whenNoDay         whenGap = "day"
+)
+
+// whenGapOf reads the request and names the first thing about its time that he
+// has not said. hasTime is whether a parser could read an instant out of it,
+// which is necessary and not sufficient: the parser answers a dayless "в 9"
+// with a day it picked.
+func whenGapOf(text string, hasTime bool) whenGap {
+	if !router.NamesAnHour(text) {
+		return whenNoHour
+	}
+	if router.NamesAnInterval(text) {
+		return whenComplete
+	}
+	if !hasTime {
+		return whenNoHour
+	}
+	if router.HourIsAmbiguous(text) {
+		return whenAmbiguousHour
+	}
+	if !router.NamesADay(text) {
+		return whenNoDay
+	}
+	return whenComplete
+}
+
+// whenQuestion is what she asks for each gap. Every one of them opens with the
+// current time, because she is reasoning from it and he cannot check that
+// reasoning unless he hears it. The hour deck varies with the attempt, like
+// every other slot; the other two say one thing and there is only one way to
+// say it.
+//
+// taken is what his last turn added, in his own words, and it goes between the
+// clock and the question (V-593). It is empty whenever his turn moved nothing
+// forward, which is the case where repeating the question verbatim is honest.
+func whenQuestion(gap whenGap, attempt int, now time.Time, taken string) (string, bool) {
+	clock := fmt.Sprintf("Сейчас %s.", now.Format("15:04"))
+	if taken != "" {
+		clock += " " + taken
+	}
+	switch gap {
+	case whenNoHour:
+		q, ok := clarifyQuestionFor(dialogue.SlotTime, attempt)
+		if !ok {
+			return "", false
+		}
+		return clock + " " + q, true
+	case whenAmbiguousHour:
+		return clock + " Это утра или вечера?", true
+	case whenNoDay:
+		return clock + " В какой день?", true
+	}
+	return "", false
+}
+
+// whenKnown — the three things he has to say about the time, and whether the
+// words so far say them. Read off the same predicates whenGapOf reads, so the
+// two cannot disagree about what is still open.
+type whenKnown struct{ hour, part, day bool }
+
+func whenKnownOf(text string, hasTime bool) whenKnown {
+	if !router.NamesAnHour(text) {
+		return whenKnown{}
+	}
+	if router.NamesAnInterval(text) {
+		return whenKnown{hour: true, part: true, day: true}
+	}
+	if !hasTime {
+		return whenKnown{}
+	}
+	return whenKnown{
+		hour: true,
+		part: !router.HourIsAmbiguous(text),
+		day:  router.NamesADay(text),
+	}
+}
+
+// movedForward reports whether b says something a did not.
+func (a whenKnown) movedForward(b whenKnown) bool {
+	return (!a.hour && b.hour) || (!a.part && b.part) || (!a.day && b.day)
+}
+
+// whenTakenLine — the acknowledgement in front of a re-ask, in the words he
+// just used (V-593).
+//
+// It is an echo and never a restatement, for the same reason the fact
+// confirmation is (V-592): a 1.7B asked to say a Russian sentence back invents.
+// Its only job is evidence that the turn between two asks was heard, so after
+// "на 9" and then "на завтра" she does not ask "утра или вечера?" twice
+// byte-identically while he wonders whether the microphone is on.
+func whenTakenLine(text string) string {
+	text = strings.TrimSpace(text)
+	text = strings.TrimRight(text, " \t.,!?;:")
+	if text == "" {
+		return ""
+	}
+	return "Поняла: " + text + "."
+}
+
+// whenTextOf is everything he has said about when, the original request plus
+// every answer he has given to a question about it.
+//
+// The answers are kept apart from the utterance on purpose. The utterance is
+// the reminder's payload, so folding "завтра" into it would have her read the
+// day back to him at the time she says it. And a time answer has to be read
+// against the request rather than alone: "завтра" names no hour, and the hour
+// it belongs to is the one she is already holding.
+func whenTextOf(q *dialogue.PendingQuestion) string {
+	if q.WhenText == "" {
+		return q.Utterance
+	}
+	return strings.TrimSpace(q.Utterance + " " + q.WhenText)
+}
+
+// slotStillMissing reports whether a slot is still open. Every slot but the
+// reminder's time is open when it is empty; the time is open until all three of
+// what he must say about it are said.
+func slotStillMissing(slot dialogue.Slot, utterance string, s dialogue.Slots) bool {
+	if len(dialogue.StillMissing([]dialogue.Slot{slot}, s)) > 0 {
+		return true
+	}
+	return slot == dialogue.SlotTime && whenGapOf(utterance, s.HasTime) != whenComplete
+}
+
+// readWhen reads the instant out of what he has said about the time, newest
+// statement first.
+//
+// The request plus his latest answer is tried before the whole history, and
+// that order is what makes a correction win: "нет, сегодня в 15:00" after "в
+// 11:00" must land on 15:00, and a parser reading left to right off the joined
+// history would find the 11 he just took back. The history is the fallback,
+// because an answer often completes an earlier one rather than replacing it -
+// "вечера" says which 9, and alone it names no hour at all.
+func (h *reactiveHandler) readWhen(ctx context.Context, intent router.Intent, q *dialogue.PendingQuestion, text string) (time.Time, bool) {
+	latest := strings.TrimSpace(q.Utterance + " " + text)
+	if router.NamesAnHour(text) {
+		if w := h.extractor.Extract(ctx, intent, latest, h.now()); w.HasTime {
+			return w.Time, true
+		}
+	}
+	if w := h.extractor.Extract(ctx, intent, whenTextOf(q), h.now()); w.HasTime {
+		return w.Time, true
+	}
+	return time.Time{}, false
+}
+
+// asksAboutTime reports whether the parked question is one about when.
+func asksAboutTime(missing []dialogue.Slot) bool {
+	for _, s := range missing {
+		if s == dialogue.SlotTime {
+			return true
+		}
+	}
+	return false
+}
+
+// stillOpen reports whether any of the slots she asked about is still unsaid.
+func stillOpen(missing []dialogue.Slot, utterance string, s dialogue.Slots) bool {
+	for _, slot := range missing {
+		if slotStillMissing(slot, utterance, s) {
+			return true
+		}
+	}
+	return false
+}
+
+// stillMissingFor is missingFor's engine, in wantedSlots order. It reads the
+// utterance as well as the slots, which plain StillMissing cannot: whether an
+// hour is ambiguous is a fact about the words, not about the instant they
+// parsed to.
+func stillMissingFor(intent router.Intent, utterance string, s dialogue.Slots) []dialogue.Slot {
+	var out []dialogue.Slot
+	for _, want := range wantedSlots[intent] {
+		if slotStillMissing(want, utterance, s) {
+			out = append(out, want)
+		}
+	}
+	return out
+}

cmd/mavend/repair.go

@@ -9,6 +9,7 @@ import (
 
 	"github.com/kami/maven/internal/lexicon"
 	"github.com/kami/maven/internal/morph"
+	"github.com/kami/maven/internal/phraser"
 	"github.com/kami/maven/internal/router"
 )
 
@@ -35,6 +36,11 @@ type routedTurn struct {
 	utterance string
 	intent    router.Intent
 	at        time.Time
+	// traceID — the persisted trace of this turn, stamped after the fact by
+	// stampLastTurn. 0 when nothing persisted, and then a spoken correction
+	// still teaches the classifier: the durable label is the half that needs a
+	// row to point at (V-636).
+	traceID int64
 }
 
 // repairWindow — how long a turn stays correctable. Long enough that he can
@@ -54,6 +60,13 @@ const repairWindow = 5 * time.Minute
 // said. The set's note in lexicon_ru_v1.json carries the same reasoning.
 var repairMarkers = lexicon.RepairMarkers()
 
+// repairNegatives — "she got it wrong" with no target. Matched against the whole
+// utterance, because these are complete sentences and the markers above are
+// fragments: "это не" needs an intent word after it, "не так поняла" does not.
+// Substring matching here would claim "не так" out of any sentence containing it
+// (V-636).
+var repairNegatives = lexicon.RepairNegatives()
+
 // repairIntents — the words he uses for each intent, as dictionary forms. They
 // used to be prefixes ("заметк"), which is what a prefix list costs: "команд"
 // also matched "командировка", and "факт" matched "фактически". morph.SameWord
@@ -147,6 +160,18 @@ func (h *reactiveHandler) recordTurn(utterance string, intent router.Intent) {
 	h.lastRouted = &routedTurn{utterance: utterance, intent: intent, at: h.now()}
 }
 
+// stampLastTurn attaches the trace id to the turn a correction would point at.
+// It cannot be done in recordTurn: the trace is written when the turn ends, and
+// recordTurn runs in the middle of it.
+func (h *reactiveHandler) stampLastTurn(utterance string, traceID int64) {
+	h.mu.Lock()
+	defer h.mu.Unlock()
+	if h.lastRouted == nil || h.lastRouted.utterance != utterance {
+		return
+	}
+	h.lastRouted.traceID = traceID
+}
+
 func (h *reactiveHandler) takeLastTurn() *routedTurn {
 	h.mu.Lock()
 	defer h.mu.Unlock()
@@ -157,6 +182,56 @@ func (h *reactiveHandler) takeLastTurn() *routedTurn {
 	return last
 }
 
+// resolveUntargetedRepair handles the cheap half of a spoken correction: he says
+// she got it wrong and does not say what it should have been (V-636).
+//
+// It is worth having on its own. V-630 made the target optional on the web for
+// the same reason: a turn marked wrong with no target is a usable negative, and
+// requiring the target would cost the correction he was willing to give. Voice
+// needs it more than the web does — naming an intent aloud means saying
+// "заметка" or "факт", which is Maven's vocabulary and not his.
+//
+// Nothing is redone and the classifier is not taught. There is no target, so
+// there is nothing to redo it as and nothing to teach. Only the label is written,
+// and she says so, because a correction he cannot see reads as one that was
+// dropped.
+func (h *reactiveHandler) resolveUntargetedRepair(ctx context.Context, text string) (string, bool) {
+	if !isRepairNegative(text) {
+		return "", false
+	}
+	last := h.takeLastTurn()
+	if last == nil || h.now().Sub(last.at) > repairWindow {
+		return "", false
+	}
+	if last.traceID == 0 {
+		// No row to point at, so there is no label to write and nothing this
+		// resolver can do. Routing the words normally is the honest outcome.
+		return "", false
+	}
+	h.labelCorrection(ctx, last, "")
+	log.Printf("voice: repair — %q marked wrong, no target given", last.utterance)
+	return phraser.A(phraser.RepairNoted, nil), true
+}
+
+// isRepairNegative matches the whole utterance, minus a leading "нет" and any
+// trailing punctuation. "нет, не так" is the shortest one he says.
+func isRepairNegative(utterance string) bool {
+	s := strings.ToLower(strings.TrimSpace(utterance))
+	s = strings.TrimRight(s, " .!?")
+	for _, p := range []string{"нет,", "нет", "no,", "no"} {
+		if rest := strings.TrimSpace(strings.TrimPrefix(s, p)); rest != s && rest != "" {
+			s = rest
+			break
+		}
+	}
+	for _, n := range repairNegatives {
+		if s == n {
+			return true
+		}
+	}
+	return false
+}
+
 // resolveRepair handles a spoken correction of the previous turn: teach the
 // classifier, redo the request under the corrected intent, and say so.
 func (h *reactiveHandler) resolveRepair(ctx context.Context, text string) (string, bool) {
@@ -182,6 +257,7 @@ func (h *reactiveHandler) resolveRepair(ctx context.Context, text string) (strin
 		learned = false
 	}
 	log.Printf("voice: repair — %q was %s, corrected to %s (learned=%v)", last.utterance, last.intent, corrected, learned)
+	h.labelCorrection(ctx, last, string(corrected))
 
 	dec := router.Decision{
 		Utterance: last.utterance,
@@ -207,3 +283,24 @@ func repairLine(say string, learned bool) string {
 	}
 	return "поняла, это " + say + " — запомнила."
 }
+
+// labelCorrection promotes a spoken correction into routing_labels, the same
+// table the /chat gesture writes (V-630, V-636).
+//
+// Two sinks and not one, because they keep different things. CorrectMisroute
+// appends a classifier seed, which is what makes the NEXT turn better today.
+// The label is what a fitted head trains on later, it survives the 14-day
+// transcript, and until now only the web produced any. A sample that only ever
+// held typed turns would skew to whatever he happens to be at a keyboard for,
+// and voice is where the hard cases are.
+//
+// Best-effort and silent. He has already been told the correction landed, and a
+// second sink failing is not his problem to hear about.
+func (h *reactiveHandler) labelCorrection(ctx context.Context, last *routedTurn, shouldBe string) {
+	if h.api == nil || last == nil || last.traceID == 0 {
+		return
+	}
+	if err := h.api.CorrectTurn(ctx, last.traceID, shouldBe); err != nil {
+		log.Printf("voice: repair: could not label trace %d: %v", last.traceID, err)
+	}
+}

cmd/mavend/repair_test.go

@@ -7,6 +7,7 @@ import (
 	"time"
 
 	"github.com/kami/maven/internal/router"
+	"github.com/kami/maven/internal/store"
 )
 
 func TestParseRepairReadsTheCorrectedIntent(t *testing.T) {
@@ -149,3 +150,95 @@ func TestRepairIntentWordCollisions(t *testing.T) {
 		}
 	}
 }
+
+// V-636. A spoken correction lands in the same table the /chat gesture writes,
+// so the sample is not limited to the turns he happened to type.
+func TestSpokenCorrectionWritesTheLabel(t *testing.T) {
+	h, st, _ := newClarifyHandler(t)
+	emb := router.NewHashEmbedder(256)
+	h.recall.embedder = emb
+	h.router = router.New(router.Config{Classifier: router.NewClassifier(emb), Extractor: h.extractor})
+	ctx := context.Background()
+
+	id, err := st.WriteRoutingTrace(ctx, store.RoutingTrace{
+		Ts: h.now(), Utterance: "купить хлеб", Intent: "fact", Source: "tap:voice",
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	h.recordTurn("купить хлеб", router.IntentFact)
+	h.stampLastTurn("купить хлеб", id)
+
+	if _, handled := h.resolveRepair(ctx, "нет, это заметка"); !handled {
+		t.Fatal("the correction was not handled")
+	}
+	labels, err := st.RoutingLabels(ctx, 5)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(labels) != 1 || labels[0].Was != "fact" || labels[0].ShouldBe != "note" {
+		t.Fatalf("labels %+v: the spoken correction did not land as a pair", labels)
+	}
+}
+
+// The cheap half, which voice needs more than the web does: naming an intent
+// aloud means saying "заметка", which is her vocabulary and not his.
+func TestUntargetedSpokenCorrection(t *testing.T) {
+	h, st, now := newClarifyHandler(t)
+	ctx := context.Background()
+	seed := func(utterance string) int64 {
+		id, err := st.WriteRoutingTrace(ctx, store.RoutingTrace{
+			Ts: h.now(), Utterance: utterance, Intent: "query", Source: "tap:voice",
+		})
+		if err != nil {
+			t.Fatal(err)
+		}
+		h.recordTurn(utterance, router.IntentQuery)
+		h.stampLastTurn(utterance, id)
+		return id
+	}
+
+	seed("поужинал")
+	reply, handled := h.resolveUntargetedRepair(ctx, "нет, не так")
+	if !handled {
+		t.Fatal("«нет, не так» was not read as a correction")
+	}
+	if reply == "" {
+		t.Error("a correction he cannot hear reads as one that was dropped")
+	}
+	labels, err := st.RoutingLabels(ctx, 5)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(labels) != 1 || labels[0].ShouldBe != "" || labels[0].Was != "query" {
+		t.Fatalf("labels %+v: want one untargeted negative naming what she chose", labels)
+	}
+
+	// Outside the window it is a fresh sentence, not a verdict.
+	seed("поужинал ещё раз")
+	*now = now.Add(repairWindow + time.Minute)
+	if _, handled := h.resolveUntargetedRepair(ctx, "не так"); handled {
+		t.Error("a correction outside the window was handled")
+	}
+}
+
+// Whole-utterance, never a substring. This is the difference between the
+// negatives and the markers, and getting it wrong would claim any sentence with
+// "не так" in it.
+func TestRepairNegativeIsTheWholeUtterance(t *testing.T) {
+	for _, s := range []string{
+		"не так поняла", "нет, не так", "ты ошиблась", "неправильно", "wrong", "no, that was wrong",
+	} {
+		if !isRepairNegative(s) {
+			t.Errorf("%q is not read as a correction", s)
+		}
+	}
+	for _, s := range []string{
+		"это не важно", "напомни не так поздно", "а не завтра", "не так, а вот так — это заметка",
+		"", "нет",
+	} {
+		if isRepairNegative(s) {
+			t.Errorf("%q was read as a correction", s)
+		}
+	}
+}

cmd/mavend/replier_llm.go

@@ -22,7 +22,7 @@ func newLLMReplier(c phraser.Completer, block func() string) *llmReplier {
 
 // Reply never fails: a clarify, a model error and an unusable generation all
 // answer from the stub, which is what keeps a turn from breaking on the model.
-func (r *llmReplier) Reply(d router.Decision) string {
+func (r *llmReplier) Reply(ctx context.Context, d router.Decision) string {
 	if d.Clarify {
 		// The deck, not the stub's single sentence: a clarify she cannot turn
 		// into a question is the line he hears most often when she misses him,
@@ -31,14 +31,22 @@ func (r *llmReplier) Reply(d router.Decision) string {
 		// a generation to say something this small.
 		return clarifyMissedLine(d)
 	}
-	out, err := r.p.PhraseReply(context.Background(), d)
+	if d.Intent == router.IntentFact {
+		// A captured fact is confirmed by echoing him, and the model is not
+		// asked (V-592). It has nothing to phrase FROM: replyContext hands it
+		// "записала факт: water \"drank\"", so every Russian word in the reply
+		// was the model's own invention, and on 2026-08-06 that was "Проверила,
+		// что ты выпел стакан воды" for "я выпил воды".
+		return phraser.FactAck(d.Utterance)
+	}
+	out, err := r.p.PhraseReply(ctx, d)
 	if err != nil || out == "" {
-		return r.stub.Reply(d)
+		return r.stub.Reply(ctx, d)
 	}
 	// The persona checks, on the live path (personaguard.go). A reply that
 	// leaks reasoning or calls him "вы" is worse than a flat one.
 	if _, ok := guardSpoken("reply", out); !ok {
-		return r.stub.Reply(d)
+		return r.stub.Reply(ctx, d)
 	}
 	return out
 }

cmd/mavend/replier_llm_test.go

@@ -22,7 +22,7 @@ func (s stubCompleter) Complete(_ context.Context, _ llm.Req) (string, error) {
 
 func TestLLMReplierPassesTheModelReplyThrough(t *testing.T) {
 	r := newLLMReplier(stubCompleter{out: `{"response":"записала, кофе закончился","mood":"neutral"}`}, nil)
-	got := r.Reply(router.Decision{Intent: router.IntentNote, Slots: router.Slots{Text: "кофе закончился"}})
+	got := r.Reply(context.Background(), router.Decision{Intent: router.IntentNote, Slots: router.Slots{Text: "кофе закончился"}})
 	if got != "записала, кофе закончился" {
 		t.Errorf("got %q, want %q", got, "записала, кофе закончился")
 	}
@@ -42,7 +42,7 @@ func TestLLMReplierFallsBackToStubOnEmpty(t *testing.T) {
 // the clarify deck rather than the stub's single sentence.
 func TestLLMReplierClarifyReadsTheDeck(t *testing.T) {
 	r := newLLMReplier(stubCompleter{out: "я всё поняла"}, nil)
-	got := r.Reply(router.Decision{Clarify: true, Utterance: "мгм"})
+	got := r.Reply(context.Background(), router.Decision{Clarify: true, Utterance: "мгм"})
 	if got == "я всё поняла" {
 		t.Fatal("a clarify must not be phrased by the model")
 	}
@@ -50,7 +50,7 @@ func TestLLMReplierClarifyReadsTheDeck(t *testing.T) {
 		t.Errorf("on clarify: got %q, want %q", got, want)
 	}
 	// Two different misses do not sound identical.
-	if same := r.Reply(router.Decision{Clarify: true, Utterance: "а"}); same == got {
+	if same := r.Reply(context.Background(), router.Decision{Clarify: true, Utterance: "а"}); same == got {
 		t.Log("two utterances hashed to the same line, which is allowed but should be rare")
 	}
 }
@@ -60,14 +60,14 @@ func TestLLMReplierClarifyReadsTheDeck(t *testing.T) {
 // produce, which is the same claim without pinning one wording.
 func assertAck(t *testing.T, r *llmReplier, d router.Decision, key, what string) {
 	t.Helper()
-	if got := r.Reply(d); !phraser.IsAck(key, nil, got) {
+	if got := r.Reply(context.Background(), d); !phraser.IsAck(key, nil, got) {
 		t.Errorf("on %s: got %q, want a %q line", what, got, key)
 	}
 }
 
 func assertStub(t *testing.T, r *llmReplier, d router.Decision, what string) {
 	t.Helper()
-	got, want := r.Reply(d), voice.NewStubReplier().Reply(d)
+	got, want := r.Reply(context.Background(), d), voice.NewStubReplier().Reply(context.Background(), d)
 	if got != want {
 		t.Errorf("on %s: got %q, want stub %q", what, got, want)
 	}

cmd/mavend/routingtrace.go

@@ -0,0 +1,178 @@
+// mavend/routingtrace.go — persisting the per-turn decision record (V-629).
+//
+// internal/decision keeps a 25-turn in-memory ring and persisted nothing, on the
+// argument that a turn record is read minutes later or never. The owner reversed
+// that on 06-08-2026, because the routing heads (V-546) cannot be fitted or
+// calibrated without real utterances and there is no other source of them. The
+// reversal is written down in docs/plans/21-persisting-the-routing-trace.md.
+//
+// The ring stays. It is what /trace reads, it is fast, and it is what a test that
+// wired no store still gets. This file is the second sink beside it, and it is
+// nil unless the daemon has a database — no store, no trace, no error.
+package main
+
+import (
+	"context"
+	"encoding/json"
+	"log"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/kami/maven/internal/decision"
+	"github.com/kami/maven/internal/store"
+)
+
+// traceWriter is the seam the handler persists through. store.Store satisfies
+// it. nil ⇒ the ring is the only sink, which is the pre-V-629 behaviour exactly.
+type traceWriter interface {
+	WriteRoutingTrace(ctx context.Context, tr store.RoutingTrace) (int64, error)
+}
+
+// traceSink wraps the store, or returns nil when there is none. A typed nil
+// pointer assigned straight into the interface would be non-nil and would panic
+// on the first turn, which is the classic shape of this bug.
+func traceSink(s *store.Store) traceWriter {
+	if s == nil {
+		return nil
+	}
+	return s
+}
+
+// The trace id rides the context, the same seam querysource.go uses and for the
+// same reason: handleText answers every reach through one string, and threading
+// a second value through the whole action dispatch would change a signature the
+// mic, telegram and the web all share. A caller that wants the id asks for a
+// sink; the mic path does not, and pays nothing.
+type traceIDKey struct{}
+
+type traceIDSink struct {
+	mu sync.Mutex
+	id int64
+}
+
+func (s *traceIDSink) note(id int64) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	s.id = id
+}
+
+// ID is the persisted trace for the turn, or 0 when nothing was persisted.
+func (s *traceIDSink) ID() int64 {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	return s.id
+}
+
+// withTraceIDSink returns a context that collects the persisted trace id, and
+// the sink to read after the turn has answered.
+func withTraceIDSink(ctx context.Context) (context.Context, *traceIDSink) {
+	sink := &traceIDSink{}
+	return context.WithValue(ctx, traceIDKey{}, sink), sink
+}
+
+func noteTraceID(ctx context.Context, id int64) {
+	if sink, ok := ctx.Value(traceIDKey{}).(*traceIDSink); ok {
+		sink.note(id)
+	}
+}
+
+// pruneTracesOnStart enforces retention once at wiring time. Pruning on write
+// alone is not enough: a box that goes quiet for a month keeps every row until
+// the next sixty-fourth turn, and "kept for fourteen days" would then be true
+// only of a box in daily use. Called for its effect and never blocks a start.
+func pruneTracesOnStart(s *store.Store, now time.Time) {
+	if s == nil {
+		return
+	}
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+	if err := s.PruneRoutingTraces(ctx, now.Add(-store.RoutingTraceRetention)); err != nil {
+		log.Printf("routing trace: prune on start: %v", err)
+	}
+}
+
+// persistDecision writes one finished record. It takes the same *decision.Record
+// the ring takes, so the two sinks cannot disagree about what the turn did.
+//
+// Errors are logged and swallowed. A trace is diagnostic and training data, and
+// a failed insert must never change what the owner hears.
+func (h *reactiveHandler) persistDecision(turnCtx context.Context, rec *decision.Record, src turnSource) {
+	ctx := turnCtx
+	if h.traces == nil || rec == nil || strings.TrimSpace(rec.Utterance) == "" {
+		return
+	}
+	// Detached from the turn's context, and bounded on its own. Two reasons, and
+	// the first is the one that matters: the turn is over by the time this runs,
+	// so a caller that hung up or timed out would cancel the insert, and the turn
+	// he abandoned halfway is exactly the one worth having. The second is that a
+	// write must not hold the reply, so it gets a second and no more.
+	ctx, cancel := context.WithTimeout(context.WithoutCancel(ctx), time.Second)
+	defer cancel()
+	claims, err := json.Marshal(rec.Claims)
+	if err != nil {
+		log.Printf("routing trace: marshal claims: %v", err)
+		return
+	}
+	tr := store.RoutingTrace{
+		Ts:                rec.Ts,
+		Utterance:         rec.Utterance,
+		Source:            string(src),
+		Winner:            rec.Winner,
+		Intent:            wonIntent(rec),
+		ClaimedBeforeHead: claimedBeforeHead(rec),
+		EncoderID:         h.encoderID,
+		Outcome:           wonAt(rec, decision.StageAction),
+		Claims:            claims,
+	}
+	id, err := h.traces.WriteRoutingTrace(ctx, tr)
+	if err != nil {
+		log.Printf("routing trace: write: %v", err)
+		return
+	}
+	// The id goes back to whoever asked for it, so /chat can offer a correction
+	// on the turn it is already showing (V-630). Noted on the ORIGINAL context,
+	// not the detached one above: the sink belongs to the caller's turn.
+	noteTraceID(turnCtx, id)
+	// And the spoken path, which has no reply to hang a badge on: a correction
+	// said out loud points at the previous turn, so it needs that turn's row
+	// (V-636, repair.go).
+	h.stampLastTurn(rec.Utterance, id)
+}
+
+// wonIntent — what the winning claimant made the turn. Read from the claim
+// rather than from the route, because a pre-route resolver wins without routing
+// and its intent is the honest answer to "what was this turn".
+func wonIntent(rec *decision.Record) string {
+	for _, c := range rec.Claims {
+		if c.Outcome == decision.Won && c.Intent != "" {
+			return c.Intent
+		}
+	}
+	return ""
+}
+
+// wonAt — the claimant that won at one stage. The action stage is what actually
+// produced the reply, which is a different question from what was routed: a
+// route that reached a gap and a route that ran are not the same turn.
+func wonAt(rec *decision.Record, stage string) string {
+	for _, c := range rec.Claims {
+		if c.Stage == stage && c.Outcome == decision.Won {
+			return c.Claimant
+		}
+	}
+	return ""
+}
+
+// claimedBeforeHead — a pre-route resolver or a stage-0 grammar answered, so the
+// turn teaches nothing about the classifier. Those are a large share of real
+// traffic, and fitting a head on them would fit it to the grammars rather than
+// to him. Recorded per turn rather than filtered on write, because which share
+// that is happens to be the number V-632 needs to know.
+func claimedBeforeHead(rec *decision.Record) bool {
+	stage, _, ok := strings.Cut(rec.Winner, ":")
+	if !ok {
+		return false
+	}
+	return stage == decision.StagePreRoute || stage == decision.StageZero
+}

cmd/mavend/routingtrace_test.go

@@ -0,0 +1,125 @@
+package main
+
+import (
+	"context"
+	"testing"
+
+	"github.com/kami/maven/internal/decision"
+	"github.com/kami/maven/internal/store"
+)
+
+// A real turn leaves a persisted trace, not only a ring entry. This is the whole
+// of V-629: without one there is nothing to fit the routing heads from.
+func TestTurnPersistsTrace(t *testing.T) {
+	ring := decision.NewRing()
+	h := traceHandler(t, ring)
+	h.traces = traceSink(h.dataStore)
+	h.encoderID = "hash-1024"
+
+	if reply := h.handleText(context.Background(), "web", "сколько сейчас времени"); reply == "" {
+		t.Fatal("turn produced no reply")
+	}
+	got, err := h.dataStore.RecentRoutingTraces(context.Background(), 5)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(got) != 1 {
+		t.Fatalf("persisted %d traces, want 1", len(got))
+	}
+	tr := got[0]
+	if tr.Utterance != "сколько сейчас времени" {
+		t.Errorf("utterance %q", tr.Utterance)
+	}
+	if tr.Source != string(sourceText) {
+		t.Errorf("source %q, want %q", tr.Source, sourceText)
+	}
+	// A stage-0 clock rule answers this one, so the turn teaches the classifier
+	// nothing and the trace has to say so.
+	if !tr.ClaimedBeforeHead {
+		t.Errorf("claimed_before_head false on winner %q", tr.Winner)
+	}
+	if tr.EncoderID != "hash-1024" {
+		t.Errorf("encoder_id %q", tr.EncoderID)
+	}
+	if len(tr.Claims) < 3 {
+		t.Errorf("claims %s: the losers and the never-asked are the point", tr.Claims)
+	}
+}
+
+// No store, no trace, and no panic. A typed nil pointer in the interface would
+// pass the nil check and die on the first turn.
+func TestNoStoreNoTrace(t *testing.T) {
+	ring := decision.NewRing()
+	h := traceHandler(t, ring)
+	h.traces = traceSink(nil)
+
+	if reply := h.handleText(context.Background(), "web", "сколько сейчас времени"); reply == "" {
+		t.Fatal("turn produced no reply")
+	}
+	if len(ring.Recent(5)) != 1 {
+		t.Error("the ring is still the first sink and must still hold the turn")
+	}
+}
+
+// An empty utterance writes nothing. A blank row carries no label and no
+// diagnosis, and it is his words the retention bound exists for.
+func TestEmptyUtteranceIsNotPersisted(t *testing.T) {
+	h := traceHandler(t, decision.NewRing())
+	h.traces = traceSink(h.dataStore)
+	h.persistDecision(context.Background(), &decision.Record{Utterance: "  "}, sourceText)
+	got, err := h.dataStore.RecentRoutingTraces(context.Background(), 5)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(got) != 0 {
+		t.Fatalf("persisted %d traces for a blank utterance", len(got))
+	}
+}
+
+var _ traceWriter = (*store.Store)(nil)
+
+// The trace id rides back to the caller, which is what makes a correction one
+// gesture: /chat already has the id, so saying "that was wrong" costs a button
+// and no lookup (V-630).
+func TestTurnHandsBackItsTraceID(t *testing.T) {
+	h := traceHandler(t, decision.NewRing())
+	h.traces = traceSink(h.dataStore)
+
+	ctx, sink := withTraceIDSink(context.Background())
+	if reply := h.handleText(ctx, "web", "сколько сейчас времени"); reply == "" {
+		t.Fatal("turn produced no reply")
+	}
+	id := sink.ID()
+	if id == 0 {
+		t.Fatal("no trace id came back, so /chat can offer no correction")
+	}
+	// And it names the turn that just ran, so the correction lands on the right
+	// utterance.
+	if err := h.dataStore.CorrectTurn(context.Background(), id, "query", h.now()); err != nil {
+		t.Fatal(err)
+	}
+	labels, err := h.dataStore.RoutingLabels(context.Background(), 5)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(labels) != 1 || labels[0].Utterance != "сколько сейчас времени" {
+		t.Fatalf("labels %+v, want the turn that just ran", labels)
+	}
+}
+
+// A turn nobody asked the id of costs nothing, which is the mic path.
+func TestTurnWithNoSinkStillPersists(t *testing.T) {
+	h := traceHandler(t, decision.NewRing())
+	h.traces = traceSink(h.dataStore)
+
+	if reply := h.handleText(context.Background(), "web", "сколько сейчас времени"); reply == "" {
+		t.Fatal("turn produced no reply")
+	}
+	got, err := h.dataStore.RecentRoutingTraces(context.Background(), 5)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(got) != 1 {
+		t.Fatalf("persisted %d traces, want 1", len(got))
+	}
+}

cmd/mavend/ruwords.go

@@ -17,6 +17,7 @@ import (
 	"time"
 
 	"github.com/kami/maven/internal/lexicon"
+	"github.com/kami/maven/internal/router"
 	"github.com/kami/maven/internal/say"
 )
 
@@ -73,22 +74,26 @@ func mentionsUnknownPlace(u string) bool {
 // date for a day she did not understand.
 const onlyNearDaysReply = "я считаю только сегодня, завтра, послезавтра и вчера — про другие дни пока не скажу."
 
-// dayWords — day references the calendar parser cannot resolve. A weekday name
-// or a "через …" phrase means he asked about a specific other day.
-var dayWords = []string{
-	"понедельник", "вторник", "сред", "четверг", "пятниц", "суббот", "воскресен",
-	"через", "monday", "tuesday", "wednesday", "thursday", "friday", "saturday", "sunday",
-}
-
 // mentionsUnknownDay reports whether the question names a day the calendar
 // parser could not resolve. Mirror of mentionsUnknownPlace: it exists only to
 // pick an honest reply over a confidently wrong one.
 //
 // Only called after ParseCalendarDate has already failed, so "завтра" and the
 // other words it does know never reach here.
+//
+// The weekday half was a list of STEMS matched with strings.Contains until
+// V-581 — "сред", "пятниц", "суббот". That is the hand-written Russian pattern
+// the sweep of 2026-08-04 took out, and it was wrong in the way such a pattern
+// always is: "среди", "средство" and "средний" all contain "сред", so a question
+// carrying any of them was answered with onlyNearDaysReply instead of the date.
+// Whole tokens now, and the weekday itself is router.WeekdayIndex, which reads
+// the lexicon and asks the dictionary about the case.
 func mentionsUnknownDay(u string) bool {
-	for _, w := range dayWords {
-		if strings.Contains(u, w) {
+	for _, tok := range quietTokens(u) {
+		if tok == "через" {
+			return true
+		}
+		if _, ok := router.WeekdayIndex(tok); ok {
 			return true
 		}
 	}
@@ -149,6 +154,11 @@ func hasDurationWords(u string) bool {
 // Used by the query handler when answering "когда я это сделал?"-style questions.
 func formatTime(t time.Time) string {
 	now := time.Now()
+	// The argument is a fact's Ts, which the store hands back as UTC. Only the
+	// last branch names a wall clock, and it named the store's until V-614: an
+	// answer to "когда я это сделал?" read hours off, in the same sentence
+	// shape the plan reads a day in.
+	t = t.Local()
 	if t.After(now.Add(-2*time.Minute)) && t.Before(now.Add(2*time.Minute)) {
 		return "только что"
 	}

cmd/mavend/ruwords_test.go

@@ -0,0 +1,57 @@
+package main
+
+import (
+	"strings"
+	"testing"
+	"time"
+)
+
+// "когда я это сделал?" answers off a fact's Ts, which the store hands back as
+// UTC, and the branch that names a wall clock printed it in whatever zone it
+// arrived in (V-614). The instant here is built three hours off this machine's
+// zone, so the assertion holds under TZ=UTC as well.
+func TestFormatTimeReadsHisClock(t *testing.T) {
+	_, off := time.Now().Zone()
+	away := time.FixedZone("away", off+3*60*60)
+	stored := time.Now().Add(-72 * time.Hour).In(away)
+
+	got := formatTime(stored)
+	if want := stored.Local().Format("15:04"); !strings.Contains(got, want) {
+		t.Errorf("formatTime = %q, want the hour on his clock (%s)", got, want)
+	}
+	if bad := stored.Format("15:04"); strings.Contains(got, bad) {
+		t.Errorf("formatTime = %q reads the zone the fact arrived in (%s)", got, bad)
+	}
+}
+
+// TestMentionsUnknownDayReadsWordsNotStems — the defect V-581 found. The
+// weekday half of this guard was a list of stems matched with strings.Contains,
+// so "среди", "средство" and "средний" all read as Wednesday and the question
+// was answered with onlyNearDaysReply instead of a date.
+//
+// The other half of the fix is coverage: a stem list stops at the forms whoever
+// wrote it thought of, and "воскресеньях" was not one of them.
+func TestMentionsUnknownDayReadsWordsNotStems(t *testing.T) {
+	for _, u := range []string{
+		"какое число в понедельник",
+		"какое число в среду",
+		"какое число в среде",
+		"что там по воскресеньям",
+		"what is the date on friday",
+		"какое число через неделю",
+	} {
+		if !mentionsUnknownDay(u) {
+			t.Errorf("mentionsUnknownDay(%q) = false, want true", u)
+		}
+	}
+	for _, u := range []string{
+		"какое число в среднем",
+		"сколько это в среднем",
+		"какое сегодня средство",
+		"какое число",
+	} {
+		if mentionsUnknownDay(u) {
+			t.Errorf("mentionsUnknownDay(%q) = true; it names no day", u)
+		}
+	}
+}

cmd/mavend/simulator_test.go

@@ -474,7 +474,7 @@ func newSimWorld(t *testing.T, sc scenario) *simWorld {
 	// used to be built on a nil API, which meant any scenario that produced an
 	// act panicked the moment the matcher was consulted.
 	matcher := tool.NewMatcher(api)
-	rtr := buildRouter(emb, matcher, config.DefaultRouterThreshold, router.NewLLMRouter(scripted))
+	rtr := buildRouter(emb, matcher, config.DefaultRouterThreshold, router.NewLLMRouter(scripted), nil)
 
 	w.handler = &reactiveHandler{
 		stt:    simTranscriber{},

cmd/mavend/sttseam_test.go

@@ -0,0 +1,43 @@
+package main
+
+import (
+	"testing"
+
+	"github.com/kami/maven/internal/config"
+	"github.com/kami/maven/internal/stt"
+)
+
+// A box with no workstation.stt block transcribes exactly as it did before the
+// seam existed: the floor is handed back untouched, and nothing probes.
+func TestSttSeamWithNoBlockIsTheFloor(t *testing.T) {
+	floor := stt.NewStub()
+	got, pair := sttSeam(&config.Config{}, floor)
+	if pair != nil {
+		t.Fatal("no block must build no pair")
+	}
+	if got != stt.Transcriber(floor) {
+		t.Fatal("no block must hand back the floor itself")
+	}
+}
+
+func TestSttSeamPrefersTheWorkstation(t *testing.T) {
+	cfg := &config.Config{Workstation: &config.WorkstationConfig{
+		URL: "http://127.0.0.1:1",
+		Stt: &config.WorkstationSttConfig{
+			URL:    "http://127.0.0.1:2/transcribe",
+			Health: "http://127.0.0.1:2/health",
+		},
+	}}
+	got, pair := sttSeam(cfg, stt.NewStub())
+	if pair == nil {
+		t.Fatal("a configured block must build a pair")
+	}
+	defer pair.Stop()
+	if got != stt.Transcriber(pair) {
+		t.Fatal("the pair is what callers must transcribe through")
+	}
+	// Nothing answers on port 2, so the seam is the floor until it does.
+	if pair.Available() {
+		t.Fatal("an unreachable workstation must not be available")
+	}
+}

cmd/mavend/telegramintake.go

@@ -0,0 +1,59 @@
+// mavend/telegramintake.go — wiring the inbound telegram poller (V-637).
+//
+// The poller reaches the daemon through ipc.CoreAPI and nothing else, so a
+// telegram turn takes exactly the path the web's POST /api/chat takes: Chat
+// returns the reply and the persisted trace id, and CorrectTurn writes the
+// label. Nothing in internal/delivery knows what a handler is.
+package main
+
+import (
+	"context"
+	"log"
+	"sync"
+
+	"github.com/kami/maven/internal/config"
+	"github.com/kami/maven/internal/delivery/telegramsink"
+	"github.com/kami/maven/internal/ipc"
+)
+
+// wireTelegramIntake starts the poller, or returns having done nothing. It is
+// nil-safe in every argument, because it is called from both boot paths — the
+// unlocked start and the passkey unlock — and telegram must behave the same on
+// either.
+//
+// A sink that will not build is logged rather than fatal here. The push half
+// already failed the boot in wireDispatcher for the same config, so a second
+// hard failure would only lose that message.
+func wireTelegramIntake(ctx context.Context, wg *sync.WaitGroup, api ipc.CoreAPI, cfg *config.Config) {
+	if cfg == nil || cfg.Telegram == nil || !cfg.Telegram.Intake || api == nil {
+		return
+	}
+	sink, err := telegramsink.New(*cfg.Telegram)
+	if err != nil {
+		log.Printf("telegram intake: %v", err)
+		return
+	}
+	poller, err := telegramsink.NewPoller(sink, chatTurnFn(api), api.CorrectTurn)
+	if err != nil {
+		log.Printf("telegram intake: %v", err)
+		return
+	}
+	wg.Add(1)
+	go func() {
+		defer wg.Done()
+		poller.Run(ctx)
+	}()
+}
+
+// chatTurnFn adapts ipc.Chat to the poller's Turn. The trace id comes back on
+// the reply because the daemon's Chat collects it off the context (V-630), so
+// the chat can offer the same correction the web does without a second op.
+func chatTurnFn(api ipc.CoreAPI) telegramsink.Turn {
+	return func(ctx context.Context, conversation, text string) (string, int64, error) {
+		reply, err := api.Chat(ctx, conversation, text)
+		if err != nil {
+			return "", 0, err
+		}
+		return reply.Reply, reply.TraceID, nil
+	}
+}

cmd/mavend/testdata/scenarios/evening_degraded.json

@@ -24,8 +24,8 @@
       "at": "21:00",
       "note": "he speaks. The whole voice path runs: push-to-talk, the STT seam parked with the golden transcript, the real router, the real store write, the phrasing contract.",
       "audio": "ru_fact",
-      "expect_reply_contains": ["записала"],
-      "expect_reply_lacks": ["записал ", "записал,", "записал.", "милый", "ваш"],
+      "expect_reply_contains": ["записала", "выпил воды"],
+      "expect_reply_lacks": ["записал ", "записал,", "записал.", "милый", "ваш", "стакан"],
       "expect_events": ["water"]
     },
     {

cmd/mavend/testdata/scenarios/morning_missed.json

@@ -81,10 +81,10 @@
     },
     {
       "at": "08:55",
-      "note": "stating a fact writes it and says so, in the feminine. This reply comes back through the replier from the scripted model, so the persona check is against generated text rather than a constant. The masculine forms are listed with their following character — \"записал \" and \"записал,\" — because \"записала\" contains \"записал\", and the earlier check on the comma alone passed on \"записал что ты выпил воды\".",
+      "note": "stating a fact writes it and says so, in the feminine, and in his own words. The reply no longer comes from the model at all (V-592): a 1.7B asked to restate «я выпил воды» answered «Проверила, что ты выпел стакан воды», so the confirmation is now a deck frame with his sentence in it. The masculine forms are listed with their following character — \"записал \" and \"записал,\" — because \"записала\" contains \"записал\".",
       "say": "я выпил воды",
-      "expect_reply_contains": ["записала"],
-      "expect_reply_lacks": ["записал ", "записал,", "записал.", "милый"],
+      "expect_reply_contains": ["записала", "я выпил воды"],
+      "expect_reply_lacks": ["записал ", "записал,", "записал.", "милый", "стакан"],
       "expect_events": ["water"]
     },
     {

cmd/mavend/tick.go

@@ -319,10 +319,7 @@ func (t *tickLoop) stopFinishedAlarms(ctx context.Context, keys []string, state
 	if len(keys) == 0 {
 		return nil
 	}
-	byName := make(map[string]loop.Rule, len(t.rules))
-	for _, r := range t.rules {
-		byName[r.Name] = r
-	}
+	byName := t.rulesByName()
 	live := keys[:0:0]
 	for _, key := range keys {
 		outcome := ""
@@ -381,19 +378,27 @@ func (t *tickLoop) repeatableRules(keys []string) []string {
 	if len(keys) == 0 {
 		return nil
 	}
-	wired := make(map[string]bool, len(t.rules))
-	for _, r := range t.rules {
-		wired[r.Name] = true
-	}
+	wired := t.rulesByName()
 	out := keys[:0:0]
 	for _, k := range keys {
-		if wired[k] {
+		if _, ok := wired[k]; ok {
 			out = append(out, k)
 		}
 	}
 	return out
 }
 
+// rulesByName indexes the wired rule set by name, for the two lookups above
+// that only care whether a key is still wired (repeatableRules) or need the
+// rule itself (stopFinishedAlarms).
+func (t *tickLoop) rulesByName() map[string]loop.Rule {
+	byName := make(map[string]loop.Rule, len(t.rules))
+	for _, r := range t.rules {
+		byName[r.Name] = r
+	}
+	return byName
+}
+
 // cachePhrase keeps the latest phrased nudge per rule for the sev4-repeat
 // path. writing under a mutex; the repeat path reads under the same. the
 // cache is bounded by the rule count (≤ ~30 per spec) so eviction is not a

cmd/mavend/tick_api.go

@@ -23,6 +23,7 @@ type daemonAPI struct {
 	chatFn           func(ctx context.Context, conversation, text string) string
 	getMCPServers    func() []ipc.MCPServerStatus
 	getEvents        func(n int) []ipc.IntakeEvent
+	getDecisions     func(n int) []ipc.TurnDecision
 	// nexus — the identity client, nil when no nexus block is configured. It
 	// is what makes ResolveEntity answerable at all; without it the store
 	// adapter's refusal stands, and a surface that wanted an entity id says so
@@ -96,10 +97,19 @@ func (d *daemonAPI) Chat(ctx context.Context, conversation, text string) (ipc.Ch
 		return ipc.ChatReply{}, errors.New("mavend: chat not available")
 	}
 	ctx, sink := withQuerySourceSink(ctx)
+	// The trace id rides back the same way (V-630), so /chat can offer a
+	// correction on the turn it is already showing. 0 when nothing persisted.
+	ctx, traces := withTraceIDSink(ctx)
 	reply := d.chatFn(ctx, conversation, text)
-	return ipc.ChatReply{Reply: reply, Source: sink.Name()}, nil
+	return ipc.ChatReply{Reply: reply, Source: sink.Name(), TraceID: traces.ID()}, nil
 }
 
+// CorrectTurn is NOT overridden here, and that is deliberate (V-630). Every other
+// diagnostic on this type exists because the daemon holds something the store
+// cannot answer from a table. A correction is a table, so the embedded store
+// adapter is already the right answer and a second implementation here would be
+// a second place for it to drift.
+
 // MCPServers — the configured MCP servers and their health (Vikunja #251).
 // Empty, not an error, when the mcp block is absent: "not configured" is the
 // default state and the web surface renders it as such.
@@ -118,6 +128,17 @@ func (d *daemonAPI) TickTrace(ctx context.Context) (ipc.TickTrace, error) {
 	return toIPCTickTrace(*trace), nil
 }
 
+// TurnDecisions — the arbitration records of the last few turns (V-564). Nil
+// getter means voice was never wired, and that is an empty list rather than an
+// error: a box with no voice path has had no turns to arbitrate, which is not a
+// fault and renders as an empty table.
+func (d *daemonAPI) TurnDecisions(ctx context.Context, n int) ([]ipc.TurnDecision, error) {
+	if d.getDecisions == nil {
+		return nil, nil
+	}
+	return d.getDecisions(n), nil
+}
+
 func (d *daemonAPI) MorningStatus(ctx context.Context) ([]ipc.MorningRoutineStatus, error) {
 	if d.getMorningStatus == nil {
 		return nil, errors.New("mavend: morning status not available")

cmd/mavend/turnrole.go

@@ -0,0 +1,354 @@
+package main
+
+import (
+	"strings"
+	"unicode"
+
+	"github.com/kami/maven/internal/dialogue"
+	"github.com/kami/maven/internal/lexicon"
+	"github.com/kami/maven/internal/router"
+)
+
+// turnRole — what this utterance IS, relative to the action Maven is in the
+// middle of assembling. The five roles are the owner's vocabulary (Vikunja
+// #558), plus the sixth answer a resolver is allowed to give: not_applicable,
+// which hands the turn back to generic dispatch.
+//
+// It exists because the arbitration used to be ordering. The clarify resolver,
+// the confirm gate, the follow-up merge and the repair marker all ran BEFORE
+// the router, so the claimant holding conversational state decided what an
+// utterance was without asking the one component whose job that is — and on
+// 2026-08-05 "какая сейчас погода в Риме?" became the time of a reminder,
+// because the extractor found "сейчас" in it and nothing looked at the rest.
+//
+// The rule that fixes that: a routed decision which stands on its own — its own
+// intent, its own slots filled from its own words — is not an answer, whatever
+// the extractor found inside it.
+type turnRole string
+
+const (
+	roleAnswer        turnRole = "answer"         // it fills the slot she asked about
+	roleCorrection    turnRole = "correction"     // it replaces a value she already had
+	roleSideQuery     turnRole = "side_query"     // a question of its own, asked mid-flow
+	roleNewRequest    turnRole = "new_request"    // a different request entirely
+	roleAside         turnRole = "aside"          // something he stated, not an answer
+	roleCancel        turnRole = "cancel"         // call the pending action off
+	roleNotApplicable turnRole = "not_applicable" // nothing is pending; not our turn
+)
+
+// frameWords — the words that can stand around a bare slot value without adding
+// a request. Every member is a closed class from internal/lexicon: the frame
+// itself, the interrogatives, the parts of a spoken clock, the days and the
+// months. Assembled once; the sets are copies, so this cannot edit them.
+var frameWords = buildFrameWords()
+
+func buildFrameWords() map[string]bool {
+	out := make(map[string]bool)
+	add := func(list []string) {
+		for _, w := range list {
+			out[strings.ToLower(w)] = true
+		}
+	}
+	add(lexicon.SlotValueFrame())
+	add(lexicon.Interrogatives())
+	add(lexicon.PartsOfDay())
+	add(lexicon.HalfHourWords())
+	add(lexicon.DayOffsetWords())
+	for i := 0; i < 7; i++ {
+		out[lexicon.Weekday(i)] = true
+	}
+	for m := 1; m <= 12; m++ {
+		out[lexicon.MonthGenitive(m)] = true
+	}
+	for hh := 0; hh <= 23; hh++ {
+		add(strings.Fields(lexicon.HourSpoken(hh)))
+	}
+	return out
+}
+
+// cancelWords — the same, for the words that call the pending action off.
+var cancelWords = buildCancelWords()
+
+func buildCancelWords() map[string]bool {
+	out := make(map[string]bool)
+	for _, w := range lexicon.DialogueCancel() {
+		out[strings.ToLower(w)] = true
+	}
+	return out
+}
+
+// turnTokens splits an utterance the way the router's own predicates do: over
+// letters and digits, lowercased, so punctuation and a clock's colon fall out.
+func turnTokens(text string) []string {
+	return strings.FieldsFunc(strings.ToLower(text), func(r rune) bool {
+		return !unicode.IsLetter(r) && !unicode.IsDigit(r)
+	})
+}
+
+// ownContent lists the tokens of an utterance that are neither frame nor value:
+// what it is about, over and above the thing she asked for. Numbers go out
+// because a number is the commonest slot value there is, and the closed number
+// and day lexicons go out with them.
+//
+// Empty ⇒ the utterance is a slot value and nothing else, however it is dressed
+// up. That is the whole test, and it is what separates "а что если в 11:00" —
+// which is a time, hedged — from "какая сейчас погода в Риме?", which leaves
+// "погода" and "риме" behind and is therefore about something.
+func ownContent(text string) []string {
+	var out []string
+	for _, tok := range turnTokens(text) {
+		if frameWords[tok] || lexicon.IsFillerParticle(tok) {
+			continue
+		}
+		if _, ok := lexicon.Cardinal(tok); ok {
+			continue
+		}
+		if _, ok := lexicon.Ordinal(tok); ok {
+			continue
+		}
+		if isNumeric(tok) {
+			continue
+		}
+		out = append(out, tok)
+	}
+	return out
+}
+
+func isNumeric(tok string) bool {
+	for _, r := range tok {
+		if !unicode.IsDigit(r) {
+			return false
+		}
+	}
+	return tok != ""
+}
+
+// isCancel reports whether the utterance is nothing but a call-off. Every
+// content token has to be a cancel word, so "забудь" ends the exchange and
+// "забудь купить молоко" does not.
+func isCancel(text string) bool {
+	content := ownContent(text)
+	if len(content) == 0 {
+		return false
+	}
+	for _, tok := range content {
+		if !cancelWords[tok] {
+			return false
+		}
+	}
+	return true
+}
+
+// carriesOwnRequest reads the ROUTED decision for the thing that decides this:
+// does the utterance ask for something in its own right? Each intent is asked
+// the question in its own terms, because "its own slots filled from its own
+// words" means a different field for each of them.
+//
+// A query or a system question needs no further evidence — the router already
+// read a question in these words. The write intents need the verb or the slot
+// that names the request, so a bare value the router guessed a home for does
+// not count as one.
+func carriesOwnRequest(dec router.Decision, text string) bool {
+	if dec.Clarify {
+		// The router itself was unsure. An utterance she could not route is
+		// not an utterance that outranks the question in front of it.
+		return false
+	}
+	switch dec.Intent {
+	case router.IntentQuery, router.IntentSystem:
+		return true
+	case router.IntentReminder:
+		return carriesReminderVerb(text)
+	case router.IntentFact, router.IntentNote:
+		return router.CarriesCaptureVerb(text)
+	case router.IntentAct:
+		// An act that resolved to a capability is a command. One that did not
+		// is words she cannot execute anyway, so it stays an answer and gets
+		// re-asked — the same thing that happens to it today.
+		return dec.Slots.HasFn
+	default: // chat
+		return false
+	}
+}
+
+// carriesReminderVerb — "напомни" and its forms, matched over tokens. The
+// reminder verbs are a closed lexicon and are not capture verbs, so
+// CarriesCaptureVerb never sees them.
+func carriesReminderVerb(text string) bool {
+	toks := turnTokens(text)
+	for _, v := range lexicon.ReminderVerbs() {
+		for _, t := range toks {
+			if t == strings.ToLower(v) {
+				return true
+			}
+		}
+	}
+	return false
+}
+
+// isPleasantry matches the WHOLE utterance against lexicon.Pleasantries, after
+// lowercasing and dropping the punctuation a greeting carries.
+//
+// Whole utterance and not tokens. Every token rule tried here was wrong on
+// something: "вечер" answers "это утра или вечера?", "нет" answers a confirm,
+// and "спокойной" alone is not an utterance at all. A greeting is a fixed
+// phrase, so matching it as one costs nothing and claims nothing else.
+func isPleasantry(text string) bool {
+	t := strings.ToLower(strings.TrimSpace(text))
+	t = strings.Trim(t, " .,!?…")
+	t = strings.Join(strings.Fields(t), " ")
+	if t == "" {
+		return false
+	}
+	for _, p := range lexicon.Pleasantries() {
+		if t == p {
+			return true
+		}
+	}
+	return false
+}
+
+// offlineOwnRequest is the shape half of the evidence: the offline token tests,
+// which cost nothing and never depend on the model that produced the routing.
+// It is also the whole answer when there is no route to read — the classifier
+// is the failure floor and a turn must never break on the model.
+func offlineOwnRequest(text string) bool {
+	return router.IsQuestionShaped(text) || router.CarriesCaptureVerb(text) || carriesReminderVerb(text)
+}
+
+// classifyTurnRole decides what this utterance is against the pending action.
+//
+// The fast path is the first line and it is a fast path to the SAME answer, not
+// a second decision procedure: an utterance with no content of its own can
+// never be a request of its own, so it can never be anything but an answer, and
+// the route below would spend a second on the resident model to say so. Every
+// other utterance is routed first, and the role is read off the decision.
+//
+// `routed` is the turn's routing, already computed; ok is false when there was
+// none to compute (no router wired, or the route failed). A failed route falls
+// to the offline shape tests rather than breaking the turn.
+func classifyTurnRole(q *dialogue.PendingQuestion, text string, answer dialogue.Slots, routed router.Decision, ok bool) turnRole {
+	if isCancel(text) {
+		return roleCancel
+	}
+	// Two pieces of evidence, and the content gate in front of both. The shape
+	// tests are the floor and answer for free; the route is what sees a request
+	// with no shape to it — "погода в риме" asks a question and carries neither
+	// a question mark nor an interrogative, and only the router knows that.
+	//
+	// An utterance of pure frame gets one more chance, and V-577 is why. Every
+	// token of "что у меня сегодня?" is frame, so the content gate called it an
+	// answer, the parked reminder took "сегодня" for its time, and the question
+	// he asked was answered nowhere. A routed intent beats a frame match,
+	// because the frame is a hint and the route is a decision.
+	//
+	// The condition is that it fills nothing she asked about. That keeps the
+	// hedged "а что если в 11:00" an answer, which is what it is: it carries the
+	// hour, and no route saying "question" changes that. It works because the
+	// extractor no longer reads a day word as the current clock, so a sentence
+	// that names no hour now fills nothing to weigh.
+	// A pleasantry is neither (V-663). "спасибо" and "привет" fell through to
+	// roleAnswer, so a question about a reminder's DAY was re-asked at a man
+	// saying thank you, and the retry it spent was one of the three bounds
+	// meant to end the ride. It is an aside: answered as itself, the question
+	// resumed on the tail, no attempt spent, one ride counted. Placed above the
+	// content gate because "доброе утро" has content and states nothing, so
+	// neither half of the evidence below can reach it.
+	if q != nil && isPleasantry(text) {
+		return roleAside
+	}
+	own := false
+	if len(ownContent(text)) > 0 {
+		own = offlineOwnRequest(text) || (ok && carriesOwnRequest(routed, text))
+	} else if ok && fillsNothingAsked(q, answer) {
+		own = carriesOwnRequest(routed, text)
+	}
+	if !own {
+		if isAside(q, text, answer, routed, ok) {
+			return roleAside
+		}
+		if replacesFilledSlot(q, answer) {
+			return roleCorrection
+		}
+		return roleAnswer
+	}
+	if router.IsQuestionShaped(text) || (ok && (routed.Intent == router.IntentQuery || routed.Intent == router.IntentSystem)) {
+		return roleSideQuery
+	}
+	return roleNewRequest
+}
+
+// isAside reports whether the utterance is something he STATED while she was
+// waiting on a question (V-577 shape 2).
+//
+// "у меня новый ноутбук" said into a parked reminder was dropped in silence: it
+// carries no capture verb, so it is not a request of its own, and it fills no
+// slot, so it is not an answer either. Neither storing it nor saying it was
+// ignored is the one behaviour that is wrong, and it was the behaviour.
+//
+// Three conditions, and all three are needed. The route has to call it a
+// statement AND stand behind that, so a bare time is never an aside. It has to
+// fill none of what she asked about, so an answer she can use stays an answer.
+// And it has to say something, so a shrug is still a failed answer and still
+// spends a retry.
+func isAside(q *dialogue.PendingQuestion, text string, answer dialogue.Slots, routed router.Decision, ok bool) bool {
+	if !ok || q == nil {
+		return false
+	}
+	if !statesSomething(routed) {
+		return false
+	}
+	if len(ownContent(text)) == 0 {
+		return false
+	}
+	return fillsNothingAsked(q, answer)
+}
+
+// statesSomething reports whether the route is evidence that these words state
+// a thing, rather than a guess she has to interrupt a flow over.
+//
+// Two kinds of evidence, and the second one exists because the classifier floor
+// marks nearly everything Clarify. A parsed fact key comes from the
+// deterministic fact parser and not from a similarity score, so "я выпил воды"
+// is a statement on any engine. A confident note or fact is the other kind, and
+// that is the one the resident model gives for "у меня новый ноутбук".
+func statesSomething(routed router.Decision) bool {
+	switch routed.Intent {
+	case router.IntentFact:
+		return routed.Slots.HasKey || !routed.Clarify
+	case router.IntentNote:
+		return !routed.Clarify
+	}
+	return false
+}
+
+// fillsNothingAsked reports whether the utterance gave her none of what she
+// asked for. Nothing is pending counts as nothing filled.
+func fillsNothingAsked(q *dialogue.PendingQuestion, answer dialogue.Slots) bool {
+	if q == nil {
+		return true
+	}
+	return len(dialogue.StillMissing(q.Missing, answer)) == len(q.Missing)
+}
+
+// replacesFilledSlot reports whether the utterance overwrites something the
+// pending action already had, rather than filling the gap she asked about —
+// "нет, на девять" while she is waiting for the subject. Both are handled the
+// same way (dialogue.Answer already prefers the newer value), so this only
+// names the turn honestly for the log and for the decision trace V-564 adds.
+func replacesFilledSlot(q *dialogue.PendingQuestion, answer dialogue.Slots) bool {
+	if q == nil {
+		return false
+	}
+	asked := make(map[dialogue.Slot]bool, len(q.Missing))
+	for _, s := range q.Missing {
+		asked[s] = true
+	}
+	if answer.HasTime && q.Slots.HasTime && !asked[dialogue.SlotTime] && !answer.Time.Equal(q.Slots.Time) {
+		return true
+	}
+	if answer.HasKey && q.Slots.HasKey && !asked[dialogue.SlotKey] && answer.Key != q.Slots.Key {
+		return true
+	}
+	return false
+}

cmd/mavend/turnrole_test.go

@@ -0,0 +1,440 @@
+package main
+
+import (
+	"context"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/dialogue"
+	"github.com/kami/maven/internal/router"
+)
+
+// TestOwnContentSeparatesAValueFromAQuestion pins the test the whole role
+// classifier rests on: after the frame, the numbers and the closed time sets
+// come out, does anything of his own remain? A hedged time leaves nothing. A
+// question about the weather leaves the weather.
+func TestOwnContentSeparatesAValueFromAQuestion(t *testing.T) {
+	cases := []struct {
+		text string
+		own  bool
+	}{
+		{"в 11:00", false},
+		{"в семь вечера", false},
+		{"нет, в 15:00", false},
+		{"а что если в 11:00", false},
+		{"на 9", false},
+		{"а, да, прости — на 9", false},
+		{"завтра", false},
+		{"в половине восьмого", false},
+		{"какая сейчас погода в Риме?", true},
+		{"кто изобрёл телефон", true},
+		{"напомни в 11:00", true},
+		{"позвонить маме", true},
+		{"запиши что я пил воду", true},
+	}
+	for _, tc := range cases {
+		if got := len(ownContent(tc.text)) > 0; got != tc.own {
+			t.Errorf("ownContent(%q) = %v, want own content = %v", tc.text, ownContent(tc.text), tc.own)
+		}
+	}
+}
+
+// TestCancelIsTheWholeUtterance — a call-off calls the request off, and a
+// sentence that merely contains the word does not.
+func TestCancelIsTheWholeUtterance(t *testing.T) {
+	for _, yes := range []string{"отмена", "забудь", "неважно", "проехали", "cancel", "ой, отмена"} {
+		if !isCancel(yes) {
+			t.Errorf("isCancel(%q) = false, want true", yes)
+		}
+	}
+	for _, no := range []string{"забудь купить молоко", "в 11:00", "позвонить маме", ""} {
+		if isCancel(no) {
+			t.Errorf("isCancel(%q) = true, want false", no)
+		}
+	}
+}
+
+// TestTurnRoleReadsTheRoutedDecision — the inversion itself. The same utterance
+// gets a different role depending on what the router made of it, which is the
+// evidence the old guard never had.
+func TestTurnRoleReadsTheRoutedDecision(t *testing.T) {
+	q := &dialogue.PendingQuestion{
+		Intent:  dialogue.Intent(router.IntentReminder),
+		Missing: []dialogue.Slot{dialogue.SlotTime},
+	}
+	dec := func(in router.Intent, s router.Slots) router.Decision {
+		return router.Decision{Intent: in, Slots: s}
+	}
+	cases := []struct {
+		name   string
+		text   string
+		routed router.Decision
+		ok     bool
+		answer dialogue.Slots
+		want   turnRole
+	}{
+		{
+			// The measured defect. The extractor finds "сейчас" and would have
+			// closed the gap with it; the route says this is a question of its
+			// own, and the question wins.
+			name:   "a world question mid-flow is a side query",
+			text:   "какая сейчас погода в Риме?",
+			routed: dec(router.IntentQuery, router.Slots{Text: "какая сейчас погода в Риме?"}),
+			ok:     true,
+			answer: dialogue.Slots{HasTime: true, Time: time.Now()},
+			want:   roleSideQuery,
+		},
+		{
+			name:   "a hedged time is an answer even routed as a query",
+			text:   "а что если в 11:00",
+			routed: dec(router.IntentQuery, router.Slots{Text: "а что если в 11:00"}),
+			ok:     true,
+			answer: dialogue.Slots{HasTime: true, Time: time.Now()},
+			want:   roleAnswer,
+		},
+		{
+			name:   "a fresh reminder is a new request",
+			text:   "напомни завтра позвонить маме",
+			routed: dec(router.IntentReminder, router.Slots{Text: "позвонить маме", HasTime: true}),
+			ok:     true,
+			want:   roleNewRequest,
+		},
+		{
+			name:   "a capture is a new request",
+			text:   "запиши что я пил воду",
+			routed: dec(router.IntentFact, router.Slots{Key: "water", HasKey: true}),
+			ok:     true,
+			want:   roleNewRequest,
+		},
+		{
+			name:   "an act that resolved to a capability is a new request",
+			text:   "выключи свет в спальне",
+			routed: dec(router.IntentAct, router.Slots{Fn: "light_off", HasFn: true}),
+			ok:     true,
+			want:   roleNewRequest,
+		},
+		{
+			// She could not route it. An utterance she did not understand does
+			// not outrank the question in front of it.
+			name:   "a clarify decision is not a request of its own",
+			text:   "выключи свет",
+			routed: router.Decision{Intent: router.IntentAct, Slots: router.Slots{Fn: "light_off", HasFn: true}, Clarify: true},
+			ok:     true,
+			want:   roleAnswer,
+		},
+		{
+			// V-577 shape 1. Every token is frame, so the content gate called
+			// this an answer and the reminder took "сегодня" for its time. It
+			// fills nothing she asked about, so the route decides, and the route
+			// says the calendar answers it.
+			name:   "an agenda question of pure frame words is a side query",
+			text:   "что у меня сегодня?",
+			routed: dec(router.IntentQuery, router.Slots{}),
+			ok:     true,
+			want:   roleSideQuery,
+		},
+		{
+			// V-577 shape 2. Neither a slot value nor a request nor a cancel.
+			// It was dropped in silence; it is an aside, and an aside is stored
+			// and re-asked.
+			name:   "a fact stated mid-flow is an aside",
+			text:   "у меня новый ноутбук",
+			routed: dec(router.IntentNote, router.Slots{Text: "у меня новый ноутбук"}),
+			ok:     true,
+			want:   roleAside,
+		},
+		{
+			// A route she is not sure of is not evidence that he stated
+			// anything, and "позвонить маме" is the answer to the other half of
+			// a reminder.
+			name:   "an unsure note is not an aside",
+			text:   "позвонить маме",
+			routed: router.Decision{Intent: router.IntentNote, Clarify: true},
+			ok:     true,
+			want:   roleAnswer,
+		},
+		{
+			name: "a bare noun that answers nothing is still an answer",
+			text: "ага",
+			ok:   false,
+			want: roleAnswer,
+		},
+		{
+			name: "no route to read falls back to the shape",
+			text: "кто изобрёл телефон",
+			ok:   false,
+			want: roleSideQuery,
+		},
+		{
+			name: "a call-off needs no route at all",
+			text: "отмена",
+			ok:   false,
+			want: roleCancel,
+		},
+	}
+	for _, tc := range cases {
+		if got := classifyTurnRole(q, tc.text, tc.answer, tc.routed, tc.ok); got != tc.want {
+			t.Errorf("%s: classifyTurnRole(%q) = %s, want %s", tc.name, tc.text, got, tc.want)
+		}
+	}
+}
+
+// TestTurnRoleNamesACorrection — the answer overwrites a slot she was not
+// asking about. Handled like an answer, named as what it is.
+func TestTurnRoleNamesACorrection(t *testing.T) {
+	nine := time.Date(2026, 8, 6, 9, 0, 0, 0, time.UTC)
+	q := &dialogue.PendingQuestion{
+		Intent:  dialogue.Intent(router.IntentReminder),
+		Missing: []dialogue.Slot{dialogue.SlotText},
+		Slots:   dialogue.Slots{HasTime: true, Time: nine.Add(2 * time.Hour)},
+	}
+	got := classifyTurnRole(q, "нет, на 9", dialogue.Slots{HasTime: true, Time: nine}, router.Decision{}, false)
+	if got != roleCorrection {
+		t.Fatalf("role = %s, want %s", got, roleCorrection)
+	}
+}
+
+// TestRomeIsAnsweredAndTheReminderIsNotInvented — the measured failure of
+// 2026-08-05, end to end through the real cascade. "напомни позвонить маме"
+// parks the time question; the weather question that follows must not become
+// its answer and must not create a reminder for a time nobody asked for.
+//
+// V-560 got that far by DROPPING the parked request and saying so, and the
+// owner rejected the notice on sight: he did not ask to lose the reminder. So
+// the contract here is V-561's — the flow is suspended, this turn's reply ends
+// with the question coming back, and nothing says anything was let go.
+func TestRomeIsAnsweredAndTheReminderIsNotInvented(t *testing.T) {
+	ctx := context.Background()
+	h, st := newRoutingClarifyHandler(t)
+
+	if reply := h.handleText(ctx, "web", "напомни позвонить маме"); !strings.Contains(reply, "?") {
+		t.Fatalf("expected the time question, got %q", reply)
+	}
+	reply := h.handleText(ctx, "web", "какая сейчас погода в Риме?")
+	if strings.Contains(reply, "напомню") {
+		t.Fatalf("the question was eaten as the reminder's time again: %q", reply)
+	}
+	if strings.Contains(reply, clarifyDropped) {
+		t.Fatalf("a side query suspends the flow; nothing was dropped, so nothing may say so: %q", reply)
+	}
+	resumed, _ := clarifyResumedFor(dialogue.SlotTime)
+	if !strings.HasSuffix(reply, resumed) {
+		t.Fatalf("the reply must end with the resumed question %q, got %q", resumed, reply)
+	}
+	if reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour)); err != nil || len(reminders) != 0 {
+		t.Fatalf("a reminder was invented for a time nobody asked for: %v err=%v", reminders, err)
+	}
+	// Still parked, and still on its first attempt: he answered the side query,
+	// not this question, so no retry may have been spent on it.
+	q := h.clarifyStore.Get(dialogueIDFor(sourceText, "web"), h.now())
+	if q == nil {
+		t.Fatal("the parked question was dropped instead of suspended")
+	}
+	if q.Attempts != 1 {
+		t.Fatalf("the side query spent a clarify attempt: attempts = %d, want 1", q.Attempts)
+	}
+	if !strings.Contains(q.Utterance, "маме") {
+		t.Fatalf("the suspended request lost what it was about: %q", q.Utterance)
+	}
+}
+
+// TestClarifyCancelEndsTheExchange — "отмена" while she is waiting calls the
+// half-built request off, out loud, and creates nothing.
+func TestClarifyCancelEndsTheExchange(t *testing.T) {
+	ctx := context.Background()
+	h, st := newRoutingClarifyHandler(t)
+
+	if reply := h.handleText(ctx, "web", "напомни позвонить маме"); !strings.Contains(reply, "?") {
+		t.Fatalf("expected the time question, got %q", reply)
+	}
+	if reply := h.handleText(ctx, "web", "отмена"); reply != clarifyCancelled {
+		t.Fatalf("reply = %q, want %q", reply, clarifyCancelled)
+	}
+	if reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour)); err != nil || len(reminders) != 0 {
+		t.Fatalf("a cancelled request still landed: %v err=%v", reminders, err)
+	}
+	if h.clarifyStore.Get(dialogueIDFor(sourceText, "web"), h.now()) != nil {
+		t.Fatal("a cancelled exchange must leave nothing parked")
+	}
+}
+
+// TestTheTurnIsRoutedOnce — the cost bound. A turn with a question parked pays
+// for one extra route and not two: the clarify resolver and the pipeline read
+// the same memo.
+func TestTheTurnIsRoutedOnce(t *testing.T) {
+	h, _ := newRoutingClarifyHandler(t)
+	rt := h.newTurnRoute("какая сейчас погода в Риме?", h.now())
+	ctx := withTurnRoute(withDialogueID(context.Background(), voiceDialogueID), rt)
+
+	first, ok := h.routeForRole(ctx, rt.text)
+	if !ok {
+		t.Fatal("the cascade must produce a decision to classify against")
+	}
+	second, _, _, err := rt.resolve(ctx)
+	if err != nil {
+		t.Fatalf("resolve: %v", err)
+	}
+	if second.Intent != first.Intent || second.Utterance != first.Utterance {
+		t.Fatalf("the pipeline routed again and got something else: %+v vs %+v", second, first)
+	}
+}
+
+// TestASuspendedQuestionDoesNotRideForever — V-654, the measured failure of
+// 2026-08-07 (docs/evals/2026-08-07-week-of-usage-transcript.md, t=51 to t=58).
+//
+// A side query suspends the parked question, spends no attempt and restarts the
+// TTL. Nothing else bounded it, so one unfilled time slot came back on the end
+// of six consecutive unrelated replies and stopped only when a seventh turn
+// happened to read as a failed answer. Three step-asides, then she lets it go
+// and says so.
+func TestASuspendedQuestionDoesNotRideForever(t *testing.T) {
+	ctx := context.Background()
+	h, st := newRoutingClarifyHandler(t)
+	id := dialogueIDFor(sourceText, "web")
+	resumed, _ := clarifyResumedFor(dialogue.SlotTime)
+
+	if reply := h.handleText(ctx, "web", "напомни позвонить маме"); !strings.Contains(reply, "?") {
+		t.Fatalf("expected the time question, got %q", reply)
+	}
+
+	// Three questions of his own. Each one is answered as itself and each one
+	// brings the open question back, exactly as V-561 asks.
+	asides := []string{
+		"о чём мы вчера говорили?",
+		"какие у меня напоминания?",
+		"сколько времени?",
+	}
+	for i, text := range asides {
+		reply := h.handleText(ctx, "web", text)
+		if !strings.HasSuffix(reply, resumed) {
+			t.Fatalf("side query %d: the question must come back, got %q", i+1, reply)
+		}
+		if strings.Contains(reply, clarifyDropped) {
+			t.Fatalf("side query %d: nothing was let go yet, so nothing may say so: %q", i+1, reply)
+		}
+		q := h.clarifyStore.Get(id, h.now())
+		if q == nil {
+			t.Fatalf("side query %d: the question was dropped early", i+1)
+		}
+		if q.Attempts != 1 {
+			t.Fatalf("side query %d: a step-aside spent an attempt: %d", i+1, q.Attempts)
+		}
+		if q.Suspends != i+1 {
+			t.Fatalf("side query %d: suspends = %d, want %d", i+1, q.Suspends, i+1)
+		}
+	}
+
+	// The fourth. She has stepped aside as often as she is willing to, so the
+	// request goes — out loud, and without the question on the tail.
+	reply := h.handleText(ctx, "web", "что у меня сегодня?")
+	if !strings.Contains(reply, clarifyDropped) {
+		t.Fatalf("the request was let go in silence: %q", reply)
+	}
+	if strings.HasSuffix(reply, resumed) {
+		t.Fatalf("a question she has let go must not be asked again: %q", reply)
+	}
+	if h.clarifyStore.Get(id, h.now()) != nil {
+		t.Fatal("the question must be gone once she has said she let it go")
+	}
+	if reminders, err := st.DueReminders(ctx, h.now().Add(48*time.Hour)); err != nil || len(reminders) != 0 {
+		t.Fatalf("a reminder was invented for a time nobody gave: %v err=%v", reminders, err)
+	}
+}
+
+// TestAnAnsweredGapResetsTheSuspendBudget — the counter measures CONSECUTIVE
+// step-asides. He filled a gap, so the run is broken and the next question
+// starts with its full allowance: a long exchange he is engaged with must not
+// run out of patience on his behalf.
+func TestAnAnsweredGapResetsTheSuspendBudget(t *testing.T) {
+	ctx := context.Background()
+	h, _ := newRoutingClarifyHandler(t)
+	id := dialogueIDFor(sourceText, "web")
+
+	// A bare "напомни" is missing both halves, so answering the subject re-parks
+	// the request with a question about the time.
+	if reply := h.handleText(ctx, "web", "напомни"); !strings.Contains(reply, "?") {
+		t.Fatalf("expected a question, got %q", reply)
+	}
+	if reply := h.handleText(ctx, "web", "какие у меня напоминания?"); reply == "" {
+		t.Fatal("the side query must be answered as itself")
+	}
+	if q := h.clarifyStore.Get(id, h.now()); q == nil || q.Suspends != 1 {
+		t.Fatalf("the side query was not counted: %+v", q)
+	}
+	if reply := h.handleText(ctx, "web", "позвонить маме"); reply == "" {
+		t.Fatal("the answer must be consumed")
+	}
+	q := h.clarifyStore.Get(id, h.now())
+	if q == nil {
+		t.Fatal("a reminder still needs its time, so a question must be parked")
+	}
+	if q.Suspends != 0 {
+		t.Fatalf("answering a gap must reset the suspend budget: suspends = %d", q.Suspends)
+	}
+	// The ride it already took is carried across the re-park (V-663). Resetting
+	// both counters here is what let one question ride twenty-six replies.
+	if q.Rides != 1 {
+		t.Fatalf("the aside it already took was forgotten: rides = %d", q.Rides)
+	}
+}
+
+// TestTwoBoundsCannotRearmEachOther — V-663.
+//
+// MaxSuspends landed and the measurement did not move: twenty-six of 140 turns
+// carried a tail before it and twenty-six after. This is the shape it misses,
+// taken from the 2026-08-08 run, where one question rode turns 7 to 13.
+//
+// An aside spends no attempt, so MaxAttempts never reaches it. A turn that
+// reads as a failed answer zeroes Suspends, so MaxSuspends never reaches the
+// asides either. Alternating the two rearms each bound with the other's
+// traffic. Rides counts both kinds and is never reset, so it is what ends this.
+func TestTwoBoundsCannotRearmEachOther(t *testing.T) {
+	ctx := context.Background()
+	h, _ := newRoutingClarifyHandler(t)
+	id := dialogueIDFor(sourceText, "web")
+	resumed, _ := clarifyResumedFor(dialogue.SlotTime)
+
+	if reply := h.handleText(ctx, "web", "напомни позвонить маме"); !strings.Contains(reply, "?") {
+		t.Fatalf("expected the time question, got %q", reply)
+	}
+
+	// Two asides. Each one rides and neither spends an attempt.
+	for i := 0; i < 2; i++ {
+		reply := h.handleText(ctx, "web", "какие у меня напоминания?")
+		if !strings.HasSuffix(reply, resumed) {
+			t.Fatalf("aside %d: the question must come back, got %q", i+1, reply)
+		}
+	}
+	q := h.clarifyStore.Get(id, h.now())
+	if q == nil || q.Rides != 2 || q.Suspends != 2 {
+		t.Fatalf("after two asides: %+v", q)
+	}
+
+	// A pleasantry. It used to read as a failed answer, so she re-asked the
+	// question at a man saying thank you and spent an attempt doing it. Now it
+	// is an aside: answered as itself, question on the tail, one more ride.
+	reply := h.handleText(ctx, "web", "спасибо")
+	if !strings.HasSuffix(reply, resumed) {
+		t.Fatalf("a pleasantry lost the parked question: %q", reply)
+	}
+	q = h.clarifyStore.Get(id, h.now())
+	if q == nil || q.Attempts != 1 {
+		t.Fatalf("a pleasantry spent an attempt: %+v", q)
+	}
+	if q.Rides != 3 {
+		t.Fatalf("a pleasantry rode free: %+v", q)
+	}
+
+	// One more ride of any kind and the request goes, out loud.
+	reply = h.handleText(ctx, "web", "какие у меня напоминания?")
+	if !strings.Contains(reply, clarifyDropped) {
+		t.Fatalf("the question rode four asides and was let go in silence: %q", reply)
+	}
+	if strings.HasSuffix(reply, resumed) {
+		t.Fatalf("a question she has let go must not be asked again: %q", reply)
+	}
+	if h.clarifyStore.Get(id, h.now()) != nil {
+		t.Fatal("the question must be gone once she has said she let it go")
+	}
+}

cmd/mavend/turnroute.go

@@ -0,0 +1,125 @@
+package main
+
+import (
+	"context"
+	"log"
+	"sync"
+	"time"
+
+	"github.com/kami/maven/internal/dialogue"
+	"github.com/kami/maven/internal/router"
+)
+
+// turnRoute is this turn's routing, computed at most once.
+//
+// It exists because the arbitration was inverted (Vikunja #560): the clarify
+// resolver now reads the routed decision before deciding what the utterance is,
+// and the pipeline then acts on that same decision. Routing twice would cost a
+// second on the resident model and — worse — could disagree with itself, which
+// is exactly the class of bug this task is about.
+type turnRoute struct {
+	h    *reactiveHandler
+	text string
+	now  time.Time
+
+	once sync.Once
+	dec  router.Decision
+	cont bool
+	prev *dialogue.Session
+	err  error
+
+	// dropped — what she let go of this turn and must say out loud. A parked
+	// request that dies without a word leaves him thinking it landed.
+	dropped string
+
+	// resume — the parked question, re-worded, to put AFTER this turn's answer
+	// (Vikunja #561). A side query does not end the flow it interrupted, so the
+	// reply carries two acts: the answer he asked for, then the question he
+	// still owes her. Empty ⇒ nothing was suspended.
+	resume string
+	// suspended — a flow is parked underneath this turn. askClarify reads it to
+	// decide between Put (replace the top) and Push (keep the flow and stack the
+	// new question on it), because a side query that needs clarifying of its own
+	// must not overwrite the thing it interrupted.
+	suspended bool
+}
+
+type turnRouteKey struct{}
+
+func (h *reactiveHandler) newTurnRoute(text string, now time.Time) *turnRoute {
+	return &turnRoute{h: h, text: text, now: now}
+}
+
+func withTurnRoute(ctx context.Context, rt *turnRoute) context.Context {
+	return context.WithValue(ctx, turnRouteKey{}, rt)
+}
+
+// turnRouteFrom returns the turn's memo, or nil when the caller is not inside
+// runTurn — a unit test calling one resolver directly, most often.
+func turnRouteFrom(ctx context.Context) *turnRoute {
+	rt, _ := ctx.Value(turnRouteKey{}).(*turnRoute)
+	return rt
+}
+
+// resolve does the routing exactly as step 5 of runTurn does it: an elliptical
+// follow-up is answered from the previous turn, everything else goes to the
+// router. One copy of that, so the pre-route the clarify resolver reads and the
+// decision the pipeline acts on cannot drift apart.
+func (r *turnRoute) resolve(ctx context.Context) (router.Decision, bool, *dialogue.Session, error) {
+	r.once.Do(func() {
+		if r.h.dialogueSessions != nil {
+			r.prev = r.h.dialogueSessions.Get(dialogueIDOf(ctx), r.now)
+		}
+		if dec, cont := continuationDecision(r.prev, r.text, r.now); cont {
+			log.Printf("voice: continuation of %s from the previous turn", dec.Intent)
+			r.dec, r.cont = dec, true
+			return
+		}
+		if r.h.router == nil {
+			r.err = router.ErrNoIntents
+			return
+		}
+		r.dec, r.err = r.h.router.Route(ctx, r.text, r.now)
+	})
+	return r.dec, r.cont, r.prev, r.err
+}
+
+// routeForRole gives the role classifier the turn's routed decision. The second
+// return is false when there is no usable decision — no router wired, or the
+// route failed — and the classifier falls back to its offline tests then. A
+// turn must never break on the model, so the error is logged and swallowed
+// here; step 5 reads the same memo and reports it the way it always has.
+func (h *reactiveHandler) routeForRole(ctx context.Context, text string) (router.Decision, bool) {
+	rt := turnRouteFrom(ctx)
+	if rt == nil {
+		rt = h.newTurnRoute(text, h.now())
+	}
+	dec, _, _, err := rt.resolve(ctx)
+	if err != nil {
+		log.Printf("voice: role — no route to classify against (%v), falling back to the offline tests", err)
+		return router.Decision{}, false
+	}
+	return dec, true
+}
+
+// needsRoute reports whether classifying this utterance's role is worth a
+// route. It is not: an utterance with no content of its own carries no request
+// of its own, so the classifier reaches the same answer without the model. A
+// call-off is the same — it is read off a closed lexicon and nothing else.
+//
+// This is a fast path to the SAME answer and must stay one. If it ever needs a
+// rule the classifier does not have, it has become a second decision procedure
+// and it is the thing V-560 deleted.
+//
+// A question shape is the exception and V-577 is why (measured 2026-08-06).
+// "что у меня сегодня?" is an interrogative, a preposition, a particle and a day
+// word, so every token of it is frame and it left no content of its own. The
+// fast path called it an answer, the parked reminder read "сегодня" as its time,
+// and the question he asked was never answered. Asked alone the same sentence
+// routes to query at stage 0, so the route knew and was never consulted.
+func needsRoute(text string) bool {
+	if isCancel(text) {
+		return false
+	}
+	return len(ownContent(text)) > 0 || router.IsQuestionShaped(text)
+}

cmd/mavend/voice.go

@@ -53,6 +53,7 @@ import (
 
 	"github.com/kami/maven/internal/audio"
 	"github.com/kami/maven/internal/crawl"
+	"github.com/kami/maven/internal/decision"
 	"github.com/kami/maven/internal/dialogue"
 	"github.com/kami/maven/internal/ipc"
 	"github.com/kami/maven/internal/lexicon"
@@ -137,6 +138,24 @@ type reactiveHandler struct {
 	// slot per reach, not one for the box. nil ⇒ no carry-over.
 	dialogueSessions *dialogue.SessionStore
 
+	// decisions holds the last few turns' arbitration records (V-564): who
+	// claimed the turn, who lost it and who was never asked. In memory and
+	// bounded, because a turn record is read minutes later or never, and none
+	// of his words belong in a table that outlives the diagnosis. nil ⇒ nothing
+	// is recorded, which is what a test that did not ask for one gets.
+	decisions *decision.Ring
+
+	// traces persists those same records (V-629, routingtrace.go). The ring is
+	// still what /trace reads; this is the second sink, and it exists because the
+	// routing heads cannot be fitted without real utterances. nil ⇒ the ring
+	// alone, which is the behaviour every box had before 06-08-2026.
+	traces traceWriter
+
+	// encoderID names the encoder body live on this box, stored beside each
+	// trace: a fitted distance means nothing under another body. Empty ⇒ no
+	// embedder, so the classifier was the keyword floor.
+	encoderID string
+
 	// clarifyStore parks the request behind an open question she asked (see
 	// clarify.go). nil ⇒ she falls back to the canned "не поняла" reply.
 	clarifyStore *dialogue.ClarifyStore
@@ -247,7 +266,31 @@ const (
 // path wraps it in stt/tts, the text path returns it as-is.
 //
 // The ordering is load-bearing — see the step comments.
-func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSource) string {
+func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSource) (reply string) {
+	// 0. the decision record (V-564). Installed here rather than in the IPC
+	// entry point, so the mic, telegram and the web all leave the same trail —
+	// a record only the web produced would be missing exactly the turns that
+	// are hardest to reproduce. It rides the context, costs a few dozen structs
+	// on a human-rate path, and no claim site can change a route with it.
+	if h.decisions != nil {
+		var rec *decision.Record
+		ctx, rec = decision.With(ctx, text)
+		decision.Expect(ctx, decision.StagePreRoute, preRouteLadder)
+		defer func() {
+			done := rec.Finish(h.now())
+			h.decisions.Push(done)
+			h.persistDecision(ctx, done, src)
+		}()
+	}
+
+	// 0b. the turn's routing, computed at most once and shared (Vikunja #560).
+	// The clarify resolver reads it to decide what this utterance IS before
+	// claiming it, and step 5 acts on the same decision — routing twice would
+	// cost a second on the resident model and could disagree with itself.
+	now := h.now()
+	rt := h.newTurnRoute(text, now)
+	ctx = withTurnRoute(ctx, rt)
+
 	// 1. expired clarify — a question was parked but its TTL ran out, so the
 	// request behind it is gone. Say that out loud (see clarify.go) and carry
 	// on: these words are still routed as a fresh utterance below, with the
@@ -263,7 +306,7 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// 2. confirm turn — if a destructive act is parked, this utterance is its
 	// y/n answer, not a fresh command. Handled before routing so "да" doesn't
 	// get classified as some other intent.
-	if reply, handled := h.resolveConfirm(ctx, text); handled {
+	if reply, handled := h.resolveConfirm(ctx, text); notePreRoute(ctx, "confirm", handled) {
 		return withNotice(expiredNotice, reply)
 	}
 
@@ -275,15 +318,27 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// so the notice is empty here in practice. withNotice anyway: every exit
 	// from runTurn carries it, and that is what stops the next one from
 	// forgetting.
-	if reply, handled := h.resolveClarifyAnswer(ctx, text); handled {
+	if reply, handled := h.resolveClarifyAnswer(ctx, text); notePreRoute(ctx, "clarify-answer", handled) {
 		return withNotice(expiredNotice, reply)
 	}
+	// It did not claim the turn. If it let a parked request go to get out of the
+	// way, that has to be said in front of whatever these words are answered
+	// with — carried on the same notice, so every exit below keeps it.
+	expiredNotice = withNotice(expiredNotice, rt.dropped)
+
+	// 3b. and if it SUSPENDED a request instead of letting it go, the question
+	// comes back on the end of whatever these words are answered with (Vikunja
+	// #561). A deferred append rather than a call at each exit: there are eight
+	// returns between here and the replier, and the flow has to survive all of
+	// them — one that forgot would be a request parked for ever, waiting for an
+	// answer to a question he never heard asked.
+	defer func() { reply = withResumed(reply, rt.resume) }()
 
 	// 4. quiet-hours toggle — keyword match, not classifier-dependent.
 	// "тихий режим" / "quiet on" would route through the classifier
 	// unreliably (it's a command, not a free-form query), so we match it
 	// before routing. Same pattern as the confirm turn above.
-	if reply, handled := h.resolveQuietToggle(ctx, text, src); handled {
+	if reply, handled := h.resolveQuietToggle(ctx, text, src); notePreRoute(ctx, "quiet-toggle", handled) {
 		return withNotice(expiredNotice, reply)
 	}
 
@@ -291,14 +346,14 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// sent. Only handled when a pending nudge is actually inside the window
 	// (snooze.go); otherwise the words route normally, because "потом" is an
 	// ordinary word and eating every one of them would break real sentences.
-	if reply, handled := h.resolveSnooze(ctx, text, src); handled {
+	if reply, handled := h.resolveSnooze(ctx, text, src); notePreRoute(ctx, "snooze", handled) {
 		return withNotice(expiredNotice, reply)
 	}
 
 	// 4c. spoken ack — "готово" closes that same nudge as `acted`. Only the
 	// contentless form is intercepted here; "выпил воды" keeps routing and
 	// closes the nudge after its fact lands (ackFromFact, step 8b).
-	if reply, handled := h.resolveAck(ctx, text, src); handled {
+	if reply, handled := h.resolveAck(ctx, text, src); notePreRoute(ctx, "ack", handled) {
 		return withNotice(expiredNotice, reply)
 	}
 
@@ -306,7 +361,15 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// turn and names what it should have been (repair.go). Before routing,
 	// like the confirm and clarify turns: routing the correction as a fresh
 	// utterance files the correction itself instead of fixing anything.
-	if reply, handled := h.resolveRepair(ctx, text); handled {
+	if reply, handled := h.resolveRepair(ctx, text); notePreRoute(ctx, "repair", handled) {
+		return withNotice(expiredNotice, reply)
+	}
+
+	// 4d-ii. and the same correction without a target — "нет, не так" (V-636).
+	// After the targeted one, which is the narrower claim: an utterance that
+	// names an intent is answered by redoing the request, and this rung only
+	// gets the ones that name nothing.
+	if reply, handled := h.resolveUntargetedRepair(ctx, text); notePreRoute(ctx, "repair-negative", handled) {
 		return withNotice(expiredNotice, reply)
 	}
 
@@ -314,7 +377,7 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// just read (ordinal.go). Before routing, and only when a list is actually
 	// bound to the session: with nothing offered, "второй" is an ordinary word
 	// and keeps routing.
-	if reply, handled := h.resolveCandidate(ctx, text, src); handled {
+	if reply, handled := h.resolveCandidate(ctx, text, src); notePreRoute(ctx, "ordinal", handled) {
 		return withNotice(expiredNotice, reply)
 	}
 
@@ -323,22 +386,7 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// missing, so no amount of routing recovers it, and the model's guess
 	// costs seconds to obtain and is close to a coin flip. Everything else
 	// goes to the router.
-	var (
-		dec  router.Decision
-		err  error
-		prev *dialogue.Session
-	)
-	now := h.now()
-	if h.dialogueSessions != nil {
-		prev = h.dialogueSessions.Get(dialogueIDOf(ctx), now)
-	}
-	cont := false
-	if dec, cont = continuationDecision(prev, text, now); cont {
-		log.Printf("voice: continuation of %s from the previous turn", dec.Intent)
-	}
-	if !cont {
-		dec, err = h.router.Route(ctx, text, now)
-	}
+	dec, cont, prev, err := rt.resolve(ctx)
 	if err != nil {
 		// ErrNoIntents ⇒ classifier unseeded (cold boot). reply with a
 		// "still warming up" rather than a wire error.
@@ -359,21 +407,31 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// ("а завтра?" … "а послезавтра?") keeps working.
 	if h.dialogueSessions != nil {
 		if !cont {
-			dec = followUpMerge(prev, dec, now)
+			merged := followUpMerge(prev, dec, now)
+			noteMerge(ctx, dec, merged)
+			dec = merged
 		}
 		if !dec.Clarify {
 			h.rememberTurn(ctx, prev, dec, now)
 		}
 	}
 
-	// 7. clarify — she is not sure. If one named thing is missing, ask about it
-	// and park the request (clarify.go); otherwise the replier's canned reply
-	// stands.
-	if dec.Clarify {
+	// 7. clarify — something she needs is missing. If one named thing is missing,
+	// ask about it and park the request (clarify.go); otherwise the replier's
+	// canned reply stands.
+	//
+	// Not gated on dec.Clarify alone (Vikunja #557). A turn the cascade routed
+	// confidently but incompletely skipped this entirely: "напомни позвонить"
+	// reached applyAction, failed on the missing time, parked nothing, and the
+	// "в семь вечера" that followed was web-searched as a world question. A
+	// required slot that missingFor names is a gap whatever the confidence.
+	if dec.Clarify || len(missingFor(dec)) > 0 {
 		if reply := h.hexisBeforeClarify(ctx, dec); reply != "" {
 			return withNotice(expiredNotice, reply)
 		}
 		if question, asked := h.askClarify(ctx, dec); asked {
+			noteTerminal(ctx, "clarify-ask", dec.Intent,
+				"the route was below the threshold, so she asked instead of acting")
 			return withNotice(expiredNotice, question)
 		}
 	}
@@ -390,6 +448,9 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 	// the round-trip stays alive.
 	replyText := h.applyAction(ctx, dec)
 	log.Printf("voice: applyAction returned: %q", replyText)
+	// A query turn was already claimed by a source inside the chain; every other
+	// intent has no chain and no scoreboard, so the handler is the winner.
+	noteTerminal(ctx, "action-handler", dec.Intent, "")
 
 	// 8b. a fact that answers a live nudge closes it as `acted` (ack.go).
 	// Silent: the fact reply stands, she does not congratulate him for it.
@@ -397,7 +458,7 @@ func (h *reactiveHandler) runTurn(ctx context.Context, text string, src turnSour
 
 	// 9. replier — phrase the reply across the router decision.
 	if replyText == "" {
-		replyText = h.replier.Reply(dec)
+		replyText = h.replier.Reply(ctx, dec)
 	}
 	return withNotice(expiredNotice, replyText)
 }
@@ -490,6 +551,17 @@ func (h *reactiveHandler) replySystem(ctx context.Context, dec router.Decision)
 	}
 }
 
+// sessionAsTurn projects a stored session onto the dialogue.Turn shape used in
+// history lists. Shared by chatHistory and rememberTurn (clarify.go) so the
+// same session is described the same way in both places.
+func sessionAsTurn(s *dialogue.Session) dialogue.Turn {
+	return dialogue.Turn{
+		Intent: s.Intent,
+		Slots:  s.Slots,
+		Text:   s.Slots.Text,
+	}
+}
+
 // chatHistory collects dialogue turns from the session store for the current
 // conversation. Returns prior user utterances (newest last) up to a depth of
 // 4 turns. Returns nil when there's no session or no history.
@@ -503,13 +575,9 @@ func (h *reactiveHandler) chatHistory(ctx context.Context) []dialogue.Turn {
 		return nil
 	}
 	// History already includes the immediate prior turn (set by the dialogue
-	// merge at lines 373-395), plus up to 3 more from deeper history.
+	// merge in runTurn's step 6, above), plus up to 3 more from deeper history.
 	out := make([]dialogue.Turn, 0, 1+len(prev.History))
-	out = append(out, dialogue.Turn{
-		Intent: prev.Intent,
-		Slots:  prev.Slots,
-		Text:   prev.Slots.Text,
-	})
+	out = append(out, sessionAsTurn(prev))
 	out = append(out, prev.History...)
 	return out
 }

cmd/mavend/voicewire.go

@@ -11,6 +11,7 @@ import (
 	"time"
 
 	"github.com/kami/maven/internal/config"
+	"github.com/kami/maven/internal/decision"
 	"github.com/kami/maven/internal/delivery"
 	"github.com/kami/maven/internal/delivery/voicesink"
 	"github.com/kami/maven/internal/dialogue"
@@ -35,7 +36,9 @@ type voiceWiring struct {
 	sessions  *voice.Sessions
 	voiceSink delivery.Sink
 	embedder  router.Embedder
-	handler   *reactiveHandler // the reactive handler for IPC Chat
+	// heads — the routing heads, nil unless embedder.heads_path is set.
+	heads   *router.RouterHeads
+	handler *reactiveHandler // the reactive handler for IPC Chat
 	// worker clients (set when configured as Remote): closed on shutdown so
 	// mavsttd / mavttsd don't keep a stale conn into a restarting daemon.
 	sttClient *worker.Client
@@ -52,7 +55,11 @@ type voiceWiring struct {
 	// unless a `workstation` block names an address. Held here only so the
 	// prober is stopped on shutdown; callers were handed it at build time.
 	pair *llm.Pair
-	mcp  *mcpWiring
+	// sttPair — CrisperWhisper 2.0 on the workstation with mavsttd as the
+	// floor, nil unless the `workstation.stt` block names an address. Held for
+	// the same reason as pair: to stop its prober on shutdown.
+	sttPair *stt.Pair
+	mcp     *mcpWiring
 	// home — the Home Assistant client, nil unless the `smarthome` block is
 	// enabled (Vikunja #256). Its devices land in the same allowlist as every
 	// other act, so nothing else here has to know about it.
@@ -71,6 +78,9 @@ func (w *voiceWiring) close() {
 	if w.embedder != nil {
 		_ = w.embedder.Close()
 	}
+	if w.heads != nil {
+		_ = w.heads.Close()
+	}
 	if w.server != nil {
 		_ = w.server.Close()
 	}
@@ -83,6 +93,9 @@ func (w *voiceWiring) close() {
 	if w.pair != nil {
 		w.pair.Stop()
 	}
+	if w.sttPair != nil {
+		w.sttPair.Stop()
+	}
 	w.mcp.close()
 }
 
@@ -111,6 +124,7 @@ func wireVoice(cfg *config.Config, coreAPI ipc.CoreAPI, phr phraser.Phraser, mem
 	} else {
 		transcriber = stt.NewStub()
 	}
+	transcriber, w.sttPair = sttSeam(cfg, transcriber)
 	w.transcriber = transcriber
 
 	// ----- tts (Stub in-process OR Remote) -----
@@ -146,8 +160,29 @@ func wireVoice(cfg *config.Config, coreAPI ipc.CoreAPI, phr phraser.Phraser, mem
 		emb = router.NewHashEmbedder(1024)
 	}
 	w.embedder = emb
+
+	// ----- router: routing heads (only when configured, and never fatal) -----
+	// A missing or broken weights file logs and leaves w.heads nil, which is
+	// byte-for-byte the cascade that shipped before V-664. Refusing to start
+	// over a routing accelerator would trade a working box for a better one.
+	if cfg.Voice.Embedder != nil && cfg.Voice.Embedder.HeadsPath != "" {
+		h, err := router.NewRouterHeads(
+			cfg.Voice.Embedder.HeadsPath,
+			cfg.Voice.Embedder.TokenizerPath,
+		)
+		if err != nil {
+			log.Printf("voice: routing heads unavailable, cascade unchanged: %v", err)
+		} else {
+			log.Printf("voice: routing heads loaded from %s", cfg.Voice.Embedder.HeadsPath)
+			w.heads = h
+		}
+	}
+
 	repairFactVectors(dataStore, emb)
 	checkStoredEmbedder(dataStore, emb)
+	// Retention is enforced on write, which is not enough on its own: a box that
+	// goes quiet keeps every trace until the next sixty-fourth turn (V-629).
+	pruneTracesOnStart(dataStore, time.Now())
 
 	// ----- tool executor (the enabled act allowlist, store-backed) -----
 	// Config tools are the declarative bootstrap: seed them into the store as
@@ -175,7 +210,7 @@ func wireVoice(cfg *config.Config, coreAPI ipc.CoreAPI, phr phraser.Phraser, mem
 	// The LAN scanner (Vikunja #257): a read, bounded to the configured
 	// subnets and rate-limited. Off unless the `netscan` block is enabled.
 	w.netscan = wireNetScan(cfg, coreAPI)
-	matcher := tool.NewMatcher(coreAPI)
+	matcher := tool.NewMatcher(coreAPI).WithAliases(toolAliases(cfg.Voice.Tools))
 
 	// ----- weather provider (Open-Meteo when configured, Stub otherwise) -----
 	var weatherProvider weather.Provider
@@ -219,7 +254,8 @@ func wireVoice(cfg *config.Config, coreAPI ipc.CoreAPI, phr phraser.Phraser, mem
 	// against the classifier's 50.0%, at about 1s a turn instead of 30ms (see
 	// config.VoiceConfig.LLMRouter). The classifier always stays wired as the
 	// fallback, so a model error never breaks a turn.
-	rtr := buildRouter(emb, matcher, threshold, pickLLMRouter(cfg.Voice.UseLLMRouter(), hot))
+	rtr := buildRouter(emb, matcher, threshold,
+		pickLLMRouter(cfg.Voice.UseLLMRouter(), hot), w.heads)
 
 	// ----- sessions registry (shared with voicesink) -----
 	sessions := voice.NewSessions()
@@ -235,21 +271,22 @@ func wireVoice(cfg *config.Config, coreAPI ipc.CoreAPI, phr phraser.Phraser, mem
 		memStore = memory.NewInMemoryStore()
 	}
 
-	// ----- dialogue (multi-turn slot carry-over; 2-min follow-up window) -----
+	// ----- dialogue (multi-turn slot carry-over; dialogueSessionTTL follow-up window) -----
 	// Store-backed when the daemon passes a store, so a restart mid-conversation
 	// keeps the thread (Vikunja #363). Sessions past their TTL are dropped on
 	// load, never revived. Clarify's parked question stays in memory only, and
 	// that is a decision rather than an omission (Vikunja #385, docs/design.md):
 	// a restart expires it, so the thread comes back and the open question does
 	// not.
+	const dialogueSessionTTL = 2 * time.Minute
 	var dialogueSessions *dialogue.SessionStore
 	if dataStore != nil {
-		dialogueSessions = dialogue.NewPersistentSessionStore(2*time.Minute, dataStore)
+		dialogueSessions = dialogue.NewPersistentSessionStore(dialogueSessionTTL, dataStore)
 		if err := dialogueSessions.Load(context.Background(), time.Now()); err != nil {
 			log.Printf("dialogue: load saved sessions: %v", err)
 		}
 	} else {
-		dialogueSessions = dialogue.NewSessionStore(2 * time.Minute)
+		dialogueSessions = dialogue.NewSessionStore(dialogueSessionTTL)
 	}
 	clarifyStore := dialogue.NewClarifyStore(clarifyTTL)
 	timeParser := router.NewPythonDateParser()
@@ -293,7 +330,17 @@ func wireVoice(cfg *config.Config, coreAPI ipc.CoreAPI, phr phraser.Phraser, mem
 		},
 		dataStore:        dataStore,
 		dialogueSessions: dialogueSessions,
-		clarifyStore:     clarifyStore,
+		// Always on (V-564). The record is the instrument the rest of V-558 is
+		// measured with, and one that only runs when a flag is set is not there
+		// on the night the misroute happens.
+		decisions: decision.NewRing(),
+		// The second sink (V-629). Same records, persisted, because the routing
+		// heads cannot be fitted from a 25-turn ring. Nil store ⇒ ring only, and
+		// EmbedderID is the same string the vector marker uses, so a trace and a
+		// stored vector name their body the same way.
+		traces:       traceSink(dataStore),
+		encoderID:    router.EmbedderID(emb),
+		clarifyStore: clarifyStore,
 		// 0 here (unset config) ⇒ the dialogue default.
 		clarifyMaxAttempts: cfg.Voice.ClarifyMaxAttempts,
 		extractor:          router.Extractor{Time: timeParser, Acts: matcher, Facts: router.DefaultFactParser{}},
@@ -351,6 +398,44 @@ func modelSeam(cfg *config.Config, resident *llm.Client) (router.Completer, *llm
 	return pair, pair
 }
 
+// sttSeam builds the transcription seam the voice path and the meeting
+// recorder share. It is modelSeam for audio and follows the same rule.
+//
+// With no `workstation.stt` block it hands back the floor untouched, which is
+// today's deploy exactly. With one, it is an stt.Pair preferring CrisperWhisper
+// 2.0 on workpc, which scores 10.4% WER in Russian against the floor's 27.5%
+// (docs/evals/2026-08-09-crisperwhisper2-russian-wer.md).
+//
+// Only the silent half of the degradation rule applies here. A worse transcript
+// is still a turn, so there is nothing to name a gap about and the fallback is
+// never spoken. That is why stt.Pair has no TranscribeRemote.
+func sttSeam(cfg *config.Config, floor stt.Transcriber) (stt.Transcriber, *stt.Pair) {
+	if cfg.Workstation == nil || cfg.Workstation.Stt == nil {
+		return floor, nil
+	}
+	s := cfg.Workstation.Stt
+	lang := ""
+	if cfg.Voice != nil {
+		lang = cfg.Voice.Lang
+		if cfg.Voice.Stt != nil && cfg.Voice.Stt.Lang != "" {
+			lang = cfg.Voice.Stt.Lang
+		}
+	}
+	pair := stt.NewPair(
+		stt.NewHTTPTranscriber(s.URL, s.Token, lang, time.Duration(s.Timeout)),
+		floor,
+		s.Health,
+		time.Duration(s.Probe),
+	)
+	pair.Start(context.Background())
+	if s.Token == "" {
+		log.Print("voice: the workstation transcriber has no token, so anything on the LAN can post audio to it")
+	}
+	log.Printf("voice: workstation transcriber at %s, probed every %s, mavsttd as the floor",
+		s.URL, time.Duration(s.Probe))
+	return pair, pair
+}
+
 func pickLLMRouter(enabled bool, c router.Completer) *router.LLMRouter {
 	if !enabled {
 		return nil
@@ -371,11 +456,12 @@ func pickLLMRouter(enabled bool, c router.Completer) *router.LLMRouter {
 //   - The embedder is provided by wireVoice: HashEmbedder (floor) when no
 //     embedder config is present, or the ONNX multilingual model when
 //     configured — same interface, one constructor change.
-//   - 6 bootstrap examples covering the 5 intents + one compound-capture
-//     placeholder. Spec calls for ~10 per intent at production; this is the
-//     bootstrapping floor swapped by tuning the seed set later.
+//   - The classifier is floored by seedClassifier, which loads one file per
+//     intent from seedDir (models/seeds/<intent>.txt) — see seedClassifier
+//     below for the current intent list and file names.
 //   - Threshold is from voice.router_threshold config (default 0.55).
-func buildRouter(emb router.Embedder, acts router.ActMatcher, threshold float64, llmR *router.LLMRouter) *router.Router {
+func buildRouter(emb router.Embedder, acts router.ActMatcher, threshold float64,
+	llmR *router.LLMRouter, heads *router.RouterHeads) *router.Router {
 	cls := router.NewClassifier(emb)
 	seedClassifier(cls)
 	grammars := router.DefaultGrammars(acts)
@@ -386,6 +472,10 @@ func buildRouter(emb router.Embedder, acts router.ActMatcher, threshold float64,
 	grammars = append(grammars, router.AgendaQueryGrammars()...)
 	// Same reason as the agenda rules, for the feeds: "что нового в лентах?"
 	// routed system and answered "пока не умею" (Vikunja #474).
+	// After the agenda rules, which are the narrower claim, and BEFORE the feed
+	// and list rules, which are not: "что такое лента" is a definition question
+	// and the feed rule would take it on the noun alone (V-655).
+	grammars = append(grammars, router.WorldQueryGrammars()...)
 	grammars = append(grammars, router.FeedQueryGrammar())
 	// The list side of the same exposure: a phrasing with no possessive in it
 	// ("список дел") routed system and never reached queryTasks (Vikunja #467).
@@ -404,6 +494,10 @@ func buildRouter(emb router.Embedder, acts router.ActMatcher, threshold float64,
 	// board noun), and before the capture marker, which would otherwise read
 	// "убери из задач купить молоко" as a new task (Vikunja #512).
 	grammars = append(grammars, router.TaskStatusGrammar())
+	// Before the capture markers, which all need an object. A capture verb
+	// alone is a fact with no key, and the clarify path asks for it rather than
+	// letting the model invent an answer (Vikunja #557).
+	grammars = append(grammars, router.BareCaptureGrammar()...)
 	grammars = append(grammars, router.TaskCaptureGrammar())
 	// After the capture marker, so "запиши" still wins over "расскажи", and
 	// last overall because it matches on the first word alone: "расскажи про
@@ -419,6 +513,7 @@ func buildRouter(emb router.Embedder, acts router.ActMatcher, threshold float64,
 		},
 		Threshold: threshold,
 		LLM:       llmR,
+		Heads:     heads,
 	})
 }
 
@@ -452,9 +547,10 @@ func seedPath() string {
 
 // seedClassifier floors the embedded examples so the cold-boot path
 // doesn't return ErrNoIntents. Loads examples from seedDir — one file per
-// intent (act.txt, reminder.txt, fact.txt, note.txt, query.txt). When the
-// classifier can't decide it falls through to Clarify — the last-resort
-// path asks the user to rephrase rather than guessing wrong.
+// intent (act.txt, reminder.txt, fact.txt, note.txt, query.txt, chat.txt,
+// system.txt). When the classifier can't decide it falls through to
+// Clarify — the last-resort path asks the user to rephrase rather than
+// guessing wrong.
 func seedClassifier(c *router.Classifier) {
 	intents := []router.Intent{
 		router.IntentAct,
@@ -504,6 +600,20 @@ func loadSeedFile(c *router.Classifier, intent router.Intent) (int, error) {
 	return count, nil
 }
 
+// toolAliases collects the spoken phrases per tool name. Without them the act
+// matcher only ever matched the English tool name, so no Russian utterance could
+// reach a tool and every homelab act fell to proposeGap (V-633).
+func toolAliases(tools []config.ToolConfig) map[string][]string {
+	out := make(map[string][]string, len(tools))
+	for _, tc := range tools {
+		if tc.Name == "" || len(tc.Aliases) == 0 {
+			continue
+		}
+		out[tc.Name] = tc.Aliases
+	}
+	return out
+}
+
 // seedTools upserts the config-declared tools into the store as enabled. Editing
 // mavend.json is a human act, so a config tool is enabled by definition; this
 // makes the declarative config the reproducible bootstrap while the store stays

cmd/mavend/weatherq.go

@@ -7,6 +7,10 @@ package main
 import (
 	"regexp"
 	"strings"
+
+	"github.com/kami/maven/internal/lexicon"
+	"github.com/kami/maven/internal/morph"
+	"github.com/kami/maven/internal/router"
 )
 
 // isWeatherQuery returns true if the utterance is about weather.
@@ -27,16 +31,42 @@ func isWeatherQuery(u string) bool {
 // come through whole and "в 5 утра" does not.
 var weatherPlace = regexp.MustCompile(`(?i)(?:^|\s)(?:в|во|in)\s+([\p{L}-]+(?:\s+[\p{L}-]+)?)`)
 
-// weatherNonPlaces — words that follow "в" in a weather question and are not
-// cities. "какая погода в доме" is the smart-home sensor, not Open-Meteo, and
-// "тепло в комнате" is the same question about the same room.
-var weatherNonPlaces = map[string]bool{
+// weatherRooms — the rooms of the house, which are the only words in this
+// guard that belong to it. "какая погода в доме" is the smart-home sensor, not
+// Open-Meteo, and "тепло в комнате" is the same question about the same room.
+//
+// The rest of the guard used to be a third copy of three closed sets that
+// already exist in the lexicon: the weekdays, the parts of the day, and the
+// words that follow "в" without naming a place (V-581). Each copy was short in
+// its own direction — "среду" but not "среде", "утром" but not "утра", "целом"
+// but not "общем" — so the same question phrased one word differently reached
+// the geocoder as a city.
+var weatherRooms = map[string]bool{
 	"доме": true, "квартире": true, "комнате": true, "спальне": true,
 	"гостиной": true, "кухне": true, "гараже": true, "офисе": true,
-	"выходные": true, "субботу": true, "воскресенье": true, "понедельник": true,
-	"вторник": true, "среду": true, "четверг": true, "пятницу": true,
-	"обед": true, "обеде": true, "утро": true, "утром": true, "вечер": true,
-	"вечером": true, "ночь": true, "ночью": true, "целом": true, "принципе": true,
+	"обед": true, "обеде": true, "выходные": true, "выходных": true,
+}
+
+// isWeatherNonPlace reports whether the word after "в" names something other
+// than a place he could ask the weather for.
+func isWeatherNonPlace(word string) bool {
+	if weatherRooms[word] {
+		return true
+	}
+	if _, ok := router.WeekdayIndex(word); ok {
+		return true
+	}
+	for _, w := range lexicon.PartsOfDay() {
+		if word == w || morph.SameWord(word, w) {
+			return true
+		}
+	}
+	for _, w := range lexicon.NotPlaceAfterV() {
+		if word == w {
+			return true
+		}
+	}
+	return false
 }
 
 // extractWeatherLocation returns the place he named, or the configured default
@@ -63,7 +93,7 @@ func extractWeatherLocation(u, defaultLoc string) string {
 	}
 	place := strings.TrimSpace(m[1])
 	first := strings.ToLower(strings.Fields(place)[0])
-	if weatherNonPlaces[first] {
+	if isWeatherNonPlace(first) {
 		return defaultLoc
 	}
 	return place

cmd/mavend/weatherq_test.go

@@ -29,6 +29,13 @@ func TestExtractWeatherLocation(t *testing.T) {
 		// the house sensors and the day words answer elsewhere.
 		{"тепло в комнате?", "Berlin", "Berlin"},
 		{"какая погода в выходные", "Berlin", "Berlin"},
+		// The cases the three private copies of the lexicon were short by
+		// (V-581): a weekday in a case the old map did not list, a part of the
+		// day in one it did not list, and "в общем".
+		{"какая погода в среде", "Berlin", "Berlin"},
+		{"какая погода в воскресеньях", "Berlin", "Berlin"},
+		{"какая погода в понедельникам", "Berlin", "Berlin"},
+		{"какая погода в общем", "Berlin", "Berlin"},
 	}
 	for _, c := range cases {
 		if got := extractWeatherLocation(c.utterance, c.def); got != c.want {

cmd/mavgpud/gpu.go

@@ -34,22 +34,31 @@ type probe struct {
 	drmDev  string
 }
 
-// foreign lists every ROCm process that is not ours. selfPID is the supervisor's
-// llama-server child, or 0 when it is not running.
+// foreign lists every ROCm process that is not ours. self holds the pids of the
+// supervisor's own children, and a child that is not running contributes 0.
+//
+// There is more than one child since 09-08-2026. CW2 registers on the KFD like
+// any ROCm job, so a supervisor that excluded only llama-server would read its
+// own transcriber as a contender, yield the card to it, and never keep a model
+// loaded again.
 //
 // An unreadable kfd tree returns no processes and no error. That is deliberate
 // and it is the safe direction only because startVRAM also has to agree before
 // anything launches: a supervisor that cannot see the KFD never sees free VRAM
 // either, because the CPT run holding the card shows up in the drm totals.
-func (p probe) foreign(selfPID int) []gpuProc {
+func (p probe) foreign(self ...int) []gpuProc {
 	entries, err := os.ReadDir(p.kfdRoot)
 	if err != nil {
 		return nil
 	}
+	mine := make(map[int]bool, len(self))
+	for _, pid := range self {
+		mine[pid] = true
+	}
 	var out []gpuProc
 	for _, e := range entries {
 		pid, err := strconv.Atoi(e.Name())
-		if err != nil || pid == selfPID {
+		if err != nil || mine[pid] {
 			continue
 		}
 		out = append(out, gpuProc{

cmd/mavgpud/gpu_test.go

@@ -1,6 +1,7 @@
 package main
 
 import (
+	"context"
 	"net/http"
 	"net/http/httptest"
 	"net/url"
@@ -8,6 +9,7 @@ import (
 	"path/filepath"
 	"strconv"
 	"testing"
+	"time"
 )
 
 // fakeKFD builds the sysfs shape the workstation actually has: one directory
@@ -47,6 +49,24 @@ func TestForeignExcludesOurChild(t *testing.T) {
 	}
 }
 
+// The transcriber is a ROCm process on the same card, so it registers on the
+// KFD exactly like a contender does. Reading it as one is what happened on
+// 2026-08-09 while CW2 ran under its own systemd unit: mavgpud yielded, waited
+// five polls, loaded the model, yielded again, and never held it for a whole
+// minute. Excluding every child is the fix and this is the test of it.
+func TestForeignExcludesEveryChild(t *testing.T) {
+	p := probe{kfdRoot: fakeKFD(t, map[int]int64{478104: 12791693312, 999: 4096, 1001: 1717986918})}
+
+	ours := p.foreign(999, 1001)
+	if len(ours) != 1 || ours[0].PID != 478104 {
+		t.Fatalf("only the CPT run is a contender, got %+v", ours)
+	}
+	// A child that is not running reports pid 0, which must exclude nothing.
+	if got := p.foreign(999, 0); len(got) != 2 {
+		t.Errorf("a stopped child excludes nobody: got %d contenders, want 2", len(got))
+	}
+}
+
 // An empty KFD tree is the state that permits a start, so it must read as empty
 // rather than as an error the caller has to interpret.
 func TestForeignEmptyAndMissing(t *testing.T) {
@@ -81,7 +101,7 @@ func TestFreeVRAM(t *testing.T) {
 // rather than hanging or proxying into a closed port. Maven reads this endpoint
 // on a timer forever, including while the workstation is busy.
 func TestHealthAndProxyRefuseWhenNotReady(t *testing.T) {
-	s := &supervisor{run: newRunner("/bin/true", nil, "")}
+	s := &supervisor{run: newRunner("fake", "/bin/true", nil, "")}
 	h := s.handler(mustURL(t, "http://127.0.0.1:1"))
 
 	for _, path := range []string{"/health", "/v1/chat/completions"} {
@@ -101,3 +121,28 @@ func mustURL(t *testing.T, s string) *url.URL {
 	}
 	return u
 }
+
+// Yielding is all or nothing. A CPT run wants the whole card, so handing back
+// the language model while the transcriber keeps 1.6GB mapped would leave the
+// other job failing its allocation, which is the outcome yielding exists to
+// prevent.
+func TestYieldStopsEveryChild(t *testing.T) {
+	idle := "while : ; do sleep 1 ; done"
+	s := &supervisor{
+		cfg:   config{EvictAfter: 1, StopGrace: duration(2 * time.Second)},
+		probe: probe{kfdRoot: fakeKFD(t, map[int]int64{478104: 12791693312})},
+		run:   newRunner("llama-server", fakeServer(t, idle), nil, ""),
+		stt:   newRunner("cw2", fakeServer(t, idle), nil, ""),
+	}
+	for _, r := range s.children() {
+		if err := r.start(); err != nil {
+			t.Fatal(err)
+		}
+	}
+	s.tick(context.Background())
+	for _, r := range s.children() {
+		if r.running() {
+			t.Errorf("%s outlived the yield", r.name)
+		}
+	}
+}

cmd/mavgpud/main.go

@@ -10,6 +10,11 @@
 // the card. Not on demand, because a 7-14B takes tens of seconds to load and a
 // world question would be answered by a gap every time the card had been quiet.
 // Not always on, because that holds 16GB against the owner's own jobs.
+//
+// It supervises a second child since 09-08-2026, the CW2 transcriber, and for
+// one reason only: it is a ROCm process on the same card. Any GPU service the
+// owner leaves running beside this daemon reads as a contender and evicts the
+// model, so the card needs one owner rather than two neighbours.
 package main
 
 import (
@@ -36,6 +41,10 @@ type config struct {
 	// owner's business and not this daemon's schema.
 	LlamaArgs []string `json:"llama_args"`
 
+	// Stt is optional. Without it mavgpud supervises llama-server alone, which
+	// is everything it did before 09-08-2026.
+	Stt *sttConfig `json:"stt,omitempty"`
+
 	KFDRoot   string `json:"kfd_root"`
 	DRMDevice string `json:"drm_device"`
 
@@ -51,6 +60,22 @@ type config struct {
 	StartAfter int `json:"start_after_polls"`
 }
 
+// sttConfig is the CW2 transcriber, which mavgpud runs for one reason: it is a
+// ROCm process on this card. Left to its own systemd unit it registers on the
+// KFD, the supervisor reads it as a contender, and llama-server is evicted
+// within two polls and restarted five polls later, forever. That thrash was
+// observed on 2026-08-09 and it is what folded the service in here.
+//
+// Maven talks to it directly, not through this daemon. There is no proxy and no
+// idle timer: at 1.6GB it denies the card to nobody, and unloading it would only
+// send the next voice turn to the homesrv floor for no gain.
+type sttConfig struct {
+	// Addr is where the service binds, and it is read only to probe /health.
+	Addr string   `json:"addr"`
+	Bin  string   `json:"bin"`
+	Args []string `json:"args"`
+}
+
 func defaults() config {
 	return config{
 		Listen:      ":8080",
@@ -99,12 +124,18 @@ func main() {
 	}
 
 	base := "http://" + cfg.LlamaAddr
-	run := newRunner(cfg.LlamaBin, cfg.LlamaArgs, base+"/health")
+	run := newRunner("llama-server", cfg.LlamaBin, cfg.LlamaArgs, base+"/health")
 	sup := &supervisor{
 		cfg:   cfg,
 		probe: probe{kfdRoot: cfg.KFDRoot, drmDev: cfg.DRMDevice},
 		run:   run,
 	}
+	if s := cfg.Stt; s != nil {
+		if s.Bin == "" || s.Addr == "" {
+			log.Fatal("mavgpud: stt needs both bin and addr")
+		}
+		sup.stt = newRunner("cw2", s.Bin, s.Args, "http://"+s.Addr+"/health")
+	}
 	sup.touch()
 
 	ctx, cancel := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
@@ -129,13 +160,17 @@ func main() {
 	shut, done := context.WithTimeout(context.Background(), 5*time.Second)
 	defer done()
 	_ = srv.Shutdown(shut)
-	run.stop(time.Duration(cfg.StopGrace))
+	for _, r := range sup.children() {
+		r.stop(time.Duration(cfg.StopGrace))
+	}
 }
 
 type supervisor struct {
 	cfg   config
 	probe probe
 	run   *runner
+	// stt is the CW2 transcriber, or nil when the config names none.
+	stt *runner
 
 	lastReq atomic.Int64 // unix nanos of the last request Maven sent
 
@@ -198,7 +233,11 @@ func (s *supervisor) loop(ctx context.Context) {
 // allocates, so we see a contender during its startup rather than after it has
 // already failed to get the memory it wanted.
 func (s *supervisor) tick(ctx context.Context) {
-	others := s.probe.foreign(s.run.pid())
+	var pids []int
+	for _, r := range s.children() {
+		pids = append(pids, r.pid())
+	}
+	others := s.probe.foreign(pids...)
 	if len(others) > 0 {
 		s.foreignStreak++
 		s.clearStreak = 0
@@ -207,31 +246,65 @@ func (s *supervisor) tick(ctx context.Context) {
 		s.clearStreak++
 	}
 
-	if s.run.running() {
-		s.run.refreshReady(ctx)
-		switch {
-		case s.foreignStreak >= s.cfg.EvictAfter:
-			log.Printf("mavgpud: yielding the card to %s", describe(others))
-			s.run.stop(time.Duration(s.cfg.StopGrace))
-		case s.idle() > time.Duration(s.cfg.IdleTimeout):
-			log.Printf("mavgpud: idle for %s, unloading", s.idle().Round(time.Second))
-			s.run.stop(time.Duration(s.cfg.StopGrace))
+	// Yielding is all or nothing. A CPT run wants the whole card, and handing
+	// back 8GB while holding 1.6GB is the shape of a failed allocation.
+	if s.foreignStreak >= s.cfg.EvictAfter && s.anyRunning() {
+		log.Printf("mavgpud: yielding the card to %s", describe(others))
+		for _, r := range s.children() {
+			r.stop(time.Duration(s.cfg.StopGrace))
 		}
 		return
 	}
 
-	if s.clearStreak < s.cfg.StartAfter {
+	clear := s.clearStreak >= s.cfg.StartAfter
+
+	if s.run.running() {
+		s.run.refreshReady(ctx)
+		if s.idle() > time.Duration(s.cfg.IdleTimeout) {
+			log.Printf("mavgpud: idle for %s, unloading", s.idle().Round(time.Second))
+			s.run.stop(time.Duration(s.cfg.StopGrace))
+		}
+	} else if clear && s.probe.freeVRAM() >= s.cfg.MinFreeVRAM {
+		s.touch() // the idle clock starts at load, not at the last request before it
+		if err := s.run.start(); err != nil {
+			log.Printf("mavgpud: start llama-server: %v", err)
+		}
+	}
+
+	if s.stt == nil {
 		return
 	}
-	if free := s.probe.freeVRAM(); free < s.cfg.MinFreeVRAM {
+	if s.stt.running() {
+		s.stt.refreshReady(ctx)
 		return
 	}
-	s.touch() // the idle clock starts at load, not at the last request before it
-	if err := s.run.start(); err != nil {
-		log.Printf("mavgpud: start llama-server: %v", err)
+	// No VRAM precondition here, unlike llama-server. That check exists because
+	// a 12B refuses to load when the card is short, and 1.6GB fits wherever the
+	// KFD is clear. Reading free VRAM would also block the transcriber for good
+	// once the language model was resident, since it holds more than the floor.
+	if clear {
+		if err := s.stt.start(); err != nil {
+			log.Printf("mavgpud: start cw2: %v", err)
+		}
 	}
 }
 
+func (s *supervisor) children() []*runner {
+	if s.stt == nil {
+		return []*runner{s.run}
+	}
+	return []*runner{s.run, s.stt}
+}
+
+func (s *supervisor) anyRunning() bool {
+	for _, r := range s.children() {
+		if r.running() {
+			return true
+		}
+	}
+	return false
+}
+
 // describe names the contenders in the log. This log is the instrument for the
 // open question in #488: whether polling the KFD misses a job that wants the
 // card without registering there.

cmd/mavgpud/runner.go

@@ -10,14 +10,18 @@ import (
 	"time"
 )
 
-// runner owns one llama-server process. Owning it is the point of the daemon:
-// the workstation cannot keep a 7-14B resident, because that holds 16GB against
+// runner owns one GPU process. Owning it is the point of the daemon: the
+// workstation cannot keep a 7-14B resident, because that holds 16GB against
 // the owner's CPT runs, Correx and the manga-recap pipeline. So the thing that
 // stays up is this, which costs no VRAM, and the model comes and goes under it.
+//
+// There are two of them since 09-08-2026: llama-server and the CW2 transcriber.
+// name is what the log calls this one.
 type runner struct {
+	name string
 	bin  string
 	args []string
-	// ready is llama-server's own /health, which answers "is a model loaded".
+	// ready is the child's own /health, which answers "is a model loaded".
 	// Loading a 7-14B takes tens of seconds, so started is not ready.
 	readyURL string
 
@@ -32,9 +36,9 @@ type runner struct {
 	http     *http.Client
 }
 
-func newRunner(bin string, args []string, readyURL string) *runner {
+func newRunner(name, bin string, args []string, readyURL string) *runner {
 	return &runner{
-		bin: bin, args: args, readyURL: readyURL,
+		name: name, bin: bin, args: args, readyURL: readyURL,
 		http: &http.Client{Timeout: 2 * time.Second},
 	}
 }
@@ -60,7 +64,7 @@ func (r *runner) isReady() bool {
 	return r.ready
 }
 
-// start launches llama-server. It returns as soon as the process exists, not
+// start launches the child. It returns as soon as the process exists, not
 // when the model is loaded.
 func (r *runner) start() error {
 	r.mu.Lock()
@@ -76,7 +80,7 @@ func (r *runner) start() error {
 		return err
 	}
 	r.cmd, r.ready, r.yielding = cmd, false, false
-	log.Printf("mavgpud: started llama-server pid=%d", cmd.Process.Pid)
+	log.Printf("mavgpud: started %s pid=%d", r.name, cmd.Process.Pid)
 	go func() {
 		err := cmd.Wait()
 		r.mu.Lock()
@@ -84,15 +88,15 @@ func (r *runner) start() error {
 		r.cmd, r.ready, r.yielding = nil, false, false
 		r.mu.Unlock()
 		if yielded {
-			log.Printf("mavgpud: llama-server stopped, card yielded (%v)", err)
+			log.Printf("mavgpud: %s stopped, card yielded (%v)", r.name, err)
 			return
 		}
-		log.Printf("mavgpud: llama-server exited: %v", err)
+		log.Printf("mavgpud: %s exited: %v", r.name, err)
 	}()
 	return nil
 }
 
-// stop ends llama-server and waits for the VRAM to come back. SIGTERM first so
+// stop ends the child and waits for the VRAM to come back. SIGTERM first so
 // it unmaps cleanly, SIGKILL after the grace window. Returning before the
 // process is gone would let the supervisor report a free card while 14GB is
 // still mapped, which is the one lie that would make yielding useless.
@@ -117,11 +121,11 @@ func (r *runner) stop(grace time.Duration) {
 		}
 		time.Sleep(100 * time.Millisecond)
 	}
-	log.Printf("mavgpud: llama-server did not exit in %s, killing", grace)
+	log.Printf("mavgpud: %s did not exit in %s, killing", r.name, grace)
 	_ = syscall.Kill(pgid, syscall.SIGKILL)
 }
 
-// refreshReady asks llama-server whether the model is loaded. Called once per
+// refreshReady asks the child whether the model is loaded. Called once per
 // supervisor tick, never per request.
 func (r *runner) refreshReady(ctx context.Context) {
 	if !r.running() {
@@ -141,6 +145,6 @@ func (r *runner) refreshReady(ctx context.Context) {
 	r.ready = ok
 	r.mu.Unlock()
 	if ok && !was {
-		log.Printf("mavgpud: model ready")
+		log.Printf("mavgpud: %s ready", r.name)
 	}
 }

cmd/mavgpud/runner_test.go

@@ -25,7 +25,7 @@ func fakeServer(t *testing.T, body string) string {
 // status of a routine yield is identical to that of a real crash. Reading the
 // mavgpud log, the two were indistinguishable (Vikunja #491).
 func TestStopMarksTheExitAsAYield(t *testing.T) {
-	r := newRunner(fakeServer(t, "while : ; do sleep 1 ; done"), nil, "")
+	r := newRunner("fake", fakeServer(t, "while : ; do sleep 1 ; done"), nil, "")
 	if err := r.start(); err != nil {
 		t.Fatalf("start: %v", err)
 	}
@@ -49,7 +49,7 @@ func TestStopMarksTheExitAsAYield(t *testing.T) {
 // Stopping when nothing is running must not arm the flag for the next child.
 // The next exit after that would be a real crash logged as a yield.
 func TestStopWithNoChildDoesNotArmTheFlag(t *testing.T) {
-	r := newRunner("/nonexistent", nil, "")
+	r := newRunner("fake", "/nonexistent", nil, "")
 	r.stop(10 * time.Millisecond)
 	r.mu.Lock()
 	defer r.mu.Unlock()

cmd/mavpoll/main.go

@@ -6,10 +6,11 @@
 // restart-free, fail-independent — a crashing poller can't touch the store key
 // (it never had it), worst case a stale env fact until the next tick.
 //
-// Two sources, each its own provenance (the loop's rules trust source):
+// Four sources, each its own provenance (the loop's rules trust source):
 //   - netdata  → poll:netdata      resource alarms (disk/mem/cert/temp)
 //   - kuma     → poll:uptimekuma   service up/down (the source of truth for it)
 //   - zenmoney → poll:zenmoney     spending/income totals (Vikunja #125)
+//   - wireguard → infer:wg         latest handshake, the presence signal
 //
 // The zenmoney source is why the token lives HERE and not in core: the poller
 // already owns every other third-party credential, it holds no store key, and
@@ -76,27 +77,31 @@ func run(args []string) error {
 	if *netdataURL == "" && *kumaURL == "" && *wgIface == "" && *zenTokenFile == "" {
 		return fmt.Errorf("nothing to poll: set -netdata, -kuma, -wg and/or -zenmoney-token-file")
 	}
+	// A bad duration or an empty -wg-cmd used to get past start and kill the
+	// poller on the first tick — time.NewTicker panics on a non-positive
+	// interval, and pollWg indexed field 0 of an empty command. A zero -timeout
+	// is worse than a crash: http.Client reads it as "no deadline", so one
+	// wedged source stalls every other source behind it forever. Refuse all
+	// three here, where the operator sees the message.
+	if *interval <= 0 {
+		return fmt.Errorf("-interval must be positive, got %s", *interval)
+	}
+	if *timeout <= 0 {
+		return fmt.Errorf("-timeout must be positive, got %s", *timeout)
+	}
+	if *wgIface != "" && strings.TrimSpace(*wgCmd) == "" {
+		return fmt.Errorf("-wg-cmd is empty but -wg is set")
+	}
 
-	// The token is read from a file, never taken as a flag value: an argv token
-	// is visible in `ps` to every user on the box and lands in the compose file
-	// and the shell history. Read once at start — a rotated token means a
-	// restart, which is cheaper than re-reading his credential every hour.
-	var zen *zenmoney.Client
-	if *zenTokenFile != "" {
-		raw, err := os.ReadFile(*zenTokenFile)
-		if err != nil {
-			return fmt.Errorf("read zenmoney token: %w", err)
-		}
-		zen, err = zenmoney.New(strings.TrimSpace(string(raw)), *zenURL, *timeout*3)
-		if err != nil {
-			return err
-		}
+	zen, err := newZenClient(*zenTokenFile, *zenURL, *timeout)
+	if err != nil {
+		return err
 	}
 
 	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
 	defer stop()
 
-	core, err := ipc.DialWait(*socket, 60*time.Second)
+	core, err := ipc.DialWait(*socket, coreDialWait)
 	if err != nil {
 		return err
 	}
@@ -117,14 +122,46 @@ func run(args []string) error {
 	// The token is never logged, not even its length.
 	log.Printf("mavpoll: polling every %s (netdata=%q kuma=%q wg=%q zenmoney=%v every %s)",
 		*interval, *netdataURL, *kumaURL, *wgIface, zen != nil, *zenInterval)
+	p.loop(ctx, *interval)
+	return nil
+}
+
+// coreDialWait — how long to wait for core's socket at start. The poller and
+// core come up together under compose, so a cold start is a wait, not a failure.
+const coreDialWait = 60 * time.Second
+
+// zenTimeoutFactor — the zenmoney client gets a longer deadline than the other
+// sources. A diff call walks his whole transaction history, where netdata and
+// kuma answer from memory.
+const zenTimeoutFactor = 3
+
+// newZenClient builds the money client, or nil when no token file was given.
+//
+// The token is read from a file, never taken as a flag value: an argv token is
+// visible in `ps` to every user on the box and lands in the compose file and
+// the shell history. Read once at start — a rotated token means a restart,
+// which is cheaper than re-reading his credential every hour.
+func newZenClient(tokenFile, baseURL string, timeout time.Duration) (*zenmoney.Client, error) {
+	if tokenFile == "" {
+		return nil, nil
+	}
+	raw, err := os.ReadFile(tokenFile)
+	if err != nil {
+		return nil, fmt.Errorf("read zenmoney token: %w", err)
+	}
+	return zenmoney.New(strings.TrimSpace(string(raw)), baseURL, timeout*zenTimeoutFactor)
+}
+
+// loop polls until the context is cancelled.
+func (p *poller) loop(ctx context.Context, interval time.Duration) {
 	p.pollOnce(ctx) // fire immediately; don't idle a full interval on start
-	t := time.NewTicker(*interval)
+	t := time.NewTicker(interval)
 	defer t.Stop()
 	for {
 		select {
 		case <-ctx.Done():
 			log.Printf("mavpoll: bye")
-			return nil
+			return
 		case <-t.C:
 			p.pollOnce(ctx)
 		}
@@ -151,8 +188,8 @@ type poller struct {
 	zenLast  time.Time
 }
 
-// pollOnce — one sweep of both sources. A failure in one source logs and does
-// NOT abort the other: netdata being down shouldn't blind kuma and vice versa.
+// pollOnce — one sweep of every configured source. A failure in one logs and
+// does NOT abort the rest: netdata being down shouldn't blind kuma.
 func (p *poller) pollOnce(ctx context.Context) {
 	now := time.Now()
 	if p.netdataURL != "" {
@@ -244,6 +281,14 @@ func (p *poller) pollZenmoney(ctx context.Context, now time.Time) error {
 
 // ---- wireguard: latest handshake → presence signal -------------------------
 
+const (
+	// wgFactKey / wgSource — the presence signal, read by the decay in core.
+	// The source says infer because a handshake is evidence he is home, not a
+	// reading of where he is.
+	wgFactKey = "wg_handshake"
+	wgSource  = "infer:wg"
+)
+
 // pollWg reads `wg show <iface> latest-handshakes` and writes a wg_handshake
 // fact (source=infer:wg) stamped with the MOST RECENT peer handshake time — not
 // now(). Presence decays from the real handshake instant, so the fact's ts must
@@ -253,9 +298,19 @@ func (p *poller) pollZenmoney(ctx context.Context, now time.Time) error {
 // `wg show` needs CAP_NET_ADMIN; run mavpoll with the cap or set -wg-cmd "sudo wg".
 func (p *poller) pollWg(ctx context.Context) error {
 	fields := strings.Fields(p.wgCmd)
+	if len(fields) == 0 {
+		return fmt.Errorf("wg command is empty")
+	}
 	args := append(fields[1:], "show", p.wgIface, "latest-handshakes")
 	out, err := exec.CommandContext(ctx, fields[0], args...).Output()
 	if err != nil {
+		// wg says why it refused on stderr — usually a missing CAP_NET_ADMIN or
+		// an interface that does not exist. Output() drops that, leaving a log
+		// line that reads "exit status 1" and diagnoses nothing.
+		var ee *exec.ExitError
+		if errors.As(err, &ee) && len(ee.Stderr) > 0 {
+			return fmt.Errorf("run %s: %w: %s", p.wgCmd, err, strings.TrimSpace(string(ee.Stderr)))
+		}
 		return fmt.Errorf("run %s: %w", p.wgCmd, err)
 	}
 	maxTs := parseMaxHandshake(string(out))
@@ -263,20 +318,19 @@ func (p *poller) pollWg(ctx context.Context) error {
 		return nil // no peer has ever handshaked → drop out of presence
 	}
 	hs := time.Unix(maxTs, 0)
-	prev, err := p.core.LatestFactBySource(ctx, "wg_handshake", "infer:wg")
+	prev, err := p.core.LatestFactBySource(ctx, wgFactKey, wgSource)
 	if err == nil && !hs.After(prev.Ts) {
 		return nil // not newer → no churn
 	}
-	if err != nil && err != ipc.ErrNoFact && !isNoFact(err) {
-		return fmt.Errorf("read wg_handshake: %w", err)
+	if err != nil && !isNoFact(err) {
+		return fmt.Errorf("read %s: %w", wgFactKey, err)
 	}
-	if _, err := p.core.WriteFact(ctx, ipc.WriteFactReq{
-		Ts: hs, Kind: "env", Key: "wg_handshake", Value: `"up"`,
-		Source: "infer:wg", Confidence: 1.0,
-	}); err != nil {
-		return fmt.Errorf("write wg_handshake: %w", err)
+	// The ts is the handshake instant, not now(): presence decays from when he
+	// was last seen.
+	if err := p.writeFact(ctx, wgFactKey, wgSource, `"up"`, hs); err != nil {
+		return err
 	}
-	log.Printf("mavpoll: wg_handshake @ %s (infer:wg)", hs.Format(time.RFC3339))
+	log.Printf("mavpoll: %s @ %s (%s)", wgFactKey, hs.Format(time.RFC3339), wgSource)
 	return nil
 }
 
@@ -431,29 +485,53 @@ func kumaState(v float64) string {
 
 // ---- helpers ---------------------------------------------------------------
 
-// writeIfChanged writes a `facts(kind=env)` row only when val differs from the
-// latest fact for (key, source). Values are stored JSON-encoded (the store's
-// convention: `"down"`, `"critical"`), matching how rules compare f.Value.
-func (p *poller) writeIfChanged(ctx context.Context, key, source, val string, now time.Time) error {
-	jv, _ := json.Marshal(val) // string never fails to marshal
+// factConfidence — every poll is a direct reading of another service, never an
+// inference, so the fact goes in at full confidence.
+const factConfidence = 1.0
+
+// unchanged reports whether the latest fact for (key, source) already holds
+// jsonVal. A missing fact is not an error here, it is the first write.
+func (p *poller) unchanged(ctx context.Context, key, source, jsonVal string) (bool, error) {
 	prev, err := p.core.LatestFactBySource(ctx, key, source)
 	switch {
-	case err == nil && prev.Value == string(jv):
-		return nil // unchanged → no churn
-	case err != nil && err != ipc.ErrNoFact && !isNoFact(err):
-		return fmt.Errorf("read %s: %w", key, err)
+	case err == nil:
+		return prev.Value == jsonVal, nil
+	case isNoFact(err):
+		return false, nil
+	default:
+		return false, fmt.Errorf("read %s: %w", key, err)
 	}
-	_, err = p.core.WriteFact(ctx, ipc.WriteFactReq{
+}
+
+// writeFact writes one `facts(kind=env)` row. Every poll in this file lands
+// here, so the row shape is written once.
+func (p *poller) writeFact(ctx context.Context, key, source, jsonVal string, now time.Time) error {
+	_, err := p.core.WriteFact(ctx, ipc.WriteFactReq{
 		Ts:         now,
 		Kind:       "env",
 		Key:        key,
-		Value:      string(jv),
+		Value:      jsonVal,
 		Source:     source,
-		Confidence: 1.0, // a direct reading, not an inference
+		Confidence: factConfidence,
 	})
 	if err != nil {
 		return fmt.Errorf("write %s: %w", key, err)
 	}
+	return nil
+}
+
+// writeIfChanged writes only when val differs from the latest fact for
+// (key, source). Values are stored JSON-encoded (the store's convention:
+// `"down"`, `"critical"`), matching how rules compare f.Value.
+func (p *poller) writeIfChanged(ctx context.Context, key, source, val string, now time.Time) error {
+	jv, _ := json.Marshal(val) // a string never fails to marshal
+	same, err := p.unchanged(ctx, key, source, string(jv))
+	if err != nil || same {
+		return err // unchanged → no churn
+	}
+	if err := p.writeFact(ctx, key, source, string(jv), now); err != nil {
+		return err
+	}
 	log.Printf("mavpoll: %s=%s (%s)", key, val, source)
 	return nil
 }
@@ -466,18 +544,12 @@ func (p *poller) writeIfChanged(ctx context.Context, key, source, val string, no
 // The log line names the key and the source, never the figures: mavpoll's log
 // is not the place his spending ends up.
 func (p *poller) writeIfChangedRaw(ctx context.Context, key, source, jsonVal string, now time.Time) error {
-	prev, err := p.core.LatestFactBySource(ctx, key, source)
-	switch {
-	case err == nil && prev.Value == jsonVal:
-		return nil
-	case err != nil && err != ipc.ErrNoFact && !isNoFact(err):
-		return fmt.Errorf("read %s: %w", key, err)
+	same, err := p.unchanged(ctx, key, source, jsonVal)
+	if err != nil || same {
+		return err
 	}
-	if _, err := p.core.WriteFact(ctx, ipc.WriteFactReq{
-		Ts: now, Kind: "env", Key: key, Value: jsonVal,
-		Source: source, Confidence: 1.0,
-	}); err != nil {
-		return fmt.Errorf("write %s: %w", key, err)
+	if err := p.writeFact(ctx, key, source, jsonVal, now); err != nil {
+		return err
 	}
 	log.Printf("mavpoll: %s updated (%s)", key, source)
 	return nil
@@ -490,11 +562,8 @@ func (p *poller) writeIfChangedRaw(ctx context.Context, key, source, jsonVal str
 // The log line names the key only, never the figures: mavpoll's log is not the
 // place his spending ends up.
 func (p *poller) writeMoneyFact(ctx context.Context, key, jsonVal string, now time.Time) error {
-	if _, err := p.core.WriteFact(ctx, ipc.WriteFactReq{
-		Ts: now, Kind: "env", Key: key, Value: jsonVal,
-		Source: zenmoney.Source, Confidence: 1.0,
-	}); err != nil {
-		return fmt.Errorf("write %s: %w", key, err)
+	if err := p.writeFact(ctx, key, zenmoney.Source, jsonVal, now); err != nil {
+		return err
 	}
 	log.Printf("mavpoll: %s read (%s)", key, zenmoney.Source)
 	return nil
@@ -506,6 +575,15 @@ func isNoFact(err error) bool {
 	return errors.Is(err, ipc.ErrNoFact)
 }
 
+// maxBodyBytes caps what a source can make the poller hold. Kuma's whole
+// metrics page is a few hundred kilobytes, so 4 MiB is slack, not a budget.
+//
+// Hitting the cap is an error, not a shorter body. A truncated kuma page parses
+// cleanly right up to the cut, and every monitor past it reads as deleted — the
+// poller would write "unknown" over live services and the down-rule would go
+// quiet. Reading one byte past the cap is how we tell full from truncated.
+const maxBodyBytes = 4 << 20
+
 func (p *poller) get(ctx context.Context, url, basicUser string) ([]byte, error) {
 	req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
 	if err != nil {
@@ -519,12 +597,15 @@ func (p *poller) get(ctx context.Context, url, basicUser string) ([]byte, error)
 		return nil, err
 	}
 	defer resp.Body.Close()
-	body, err := io.ReadAll(io.LimitReader(resp.Body, 4<<20))
+	body, err := io.ReadAll(io.LimitReader(resp.Body, maxBodyBytes+1))
 	if err != nil {
 		return nil, err
 	}
 	if resp.StatusCode != http.StatusOK {
 		return nil, fmt.Errorf("GET %s: %s", url, resp.Status)
 	}
+	if len(body) > maxBodyBytes {
+		return nil, fmt.Errorf("GET %s: body over %d bytes", url, maxBodyBytes)
+	}
 	return body, nil
 }

cmd/mavpoll/main_test.go

@@ -3,6 +3,7 @@ package main
 import (
 	"context"
 	"encoding/json"
+	"io"
 	"net/http"
 	"net/http/httptest"
 	"os"
@@ -212,3 +213,59 @@ func TestRunRequiresSomethingToPoll(t *testing.T) {
 		t.Errorf("err = %v, want a 'nothing to poll' refusal", err)
 	}
 }
+
+// A flag value that would kill the poller later is refused at start, before it
+// dials core: a non-positive interval panics time.NewTicker on the first tick, a
+// zero timeout means http.Client waits forever, and an empty wg command used to
+// index field 0 of an empty slice.
+func TestRunRefusesFlagsThatCrashLater(t *testing.T) {
+	cases := []struct {
+		name string
+		args []string
+		want string
+	}{
+		{"zero interval", []string{"-interval", "0"}, "-interval must be positive"},
+		{"negative interval", []string{"-interval", "-5s"}, "-interval must be positive"},
+		{"zero timeout", []string{"-timeout", "0"}, "-timeout must be positive"},
+		{"empty wg command", []string{"-wg", "wg0", "-wg-cmd", " "}, "-wg-cmd is empty"},
+	}
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			args := append([]string{"-socket", "/tmp/nope.sock"}, c.args...)
+			err := run(args)
+			if err == nil || !strings.Contains(err.Error(), c.want) {
+				t.Errorf("err = %v, want %q", err, c.want)
+			}
+		})
+	}
+}
+
+// pollWg refuses an empty command rather than panicking on fields[0].
+func TestPollWgEmptyCommand(t *testing.T) {
+	p := &poller{core: &factCore{}, wgIface: "wg0", wgCmd: ""}
+	if err := p.pollWg(context.Background()); err == nil {
+		t.Error("want an error, got a poll that ran something")
+	}
+}
+
+// A body at the cap is a truncated body, and a truncated kuma page reads as
+// "every monitor past the cut was deleted". Refuse it instead of parsing it.
+func TestGetRefusesTruncatedBody(t *testing.T) {
+	big := strings.Repeat("x", maxBodyBytes+64)
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		io.WriteString(w, big)
+	}))
+	defer srv.Close()
+	p := &poller{http: srv.Client()}
+	if _, err := p.get(context.Background(), srv.URL, ""); err == nil {
+		t.Error("want an over-size refusal, got a silently truncated body")
+	}
+	small := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		io.WriteString(w, "ok")
+	}))
+	defer small.Close()
+	body, err := p.get(context.Background(), small.URL, "")
+	if err != nil || string(body) != "ok" {
+		t.Errorf("get = %q, %v; want the whole small body", body, err)
+	}
+}

cmd/mavsttd/whisper_handler.go

@@ -22,6 +22,18 @@ import (
 // to it before sending, so it's also the rate the silence gate assumes.
 const whisperSampleRate = 16000
 
+// whisperThreads — greedy decode is single-pass and this is a laptop CPU
+// (homesrv), not a server box; 4 was picked to leave headroom for the rest
+// of the daemons sharing the machine, not measured against a latency target.
+const whisperThreads = 4
+
+// noSpeechFloor — whisper's own no_speech_prob past this point means the
+// segment it transcribed is not speech (the model still emits token
+// probabilities for silence/noise, so a high avgLogProb-derived confidence
+// can coexist with a segment that should be zero). Read as "at least 90%
+// sure this was not speech."
+const noSpeechFloor = 0.9
+
 type whisperHandler struct {
 	ctx    *C.struct_whisper_context
 	minMs  int     // clips shorter than this are dropped (hallucination bait)
@@ -101,7 +113,7 @@ func (h *whisperHandler) Transcribe(ctx context.Context, req worker.TranscribeRe
 	params.print_realtime = false
 	params.print_timestamps = false
 	params.print_special = false
-	params.n_threads = C.int(4)
+	params.n_threads = C.int(whisperThreads)
 	params.single_segment = true
 
 	lang := C.CString(req.Lang)
@@ -162,7 +174,7 @@ func (h *whisperHandler) Transcribe(ctx context.Context, req worker.TranscribeRe
 	}
 
 	noSpeechProb := float64(C.whisper_full_get_segment_no_speech_prob(h.ctx, 0))
-	if noSpeechProb > 0.9 {
+	if noSpeechProb > noSpeechFloor {
 		confidence = 0
 	}
 

cmd/mavttsd/piper_handler.go

@@ -116,18 +116,27 @@ func (h *piperHandler) Synthesize(ctx context.Context, req worker.SynthesizeReq)
 	}, nil
 }
 
-// resample22050To16000 converts raw 16-bit PCM from 22050 Hz to 16000 Hz
-// using linear interpolation.
+// piperSampleRate is the rate piper's onnx voices render at (ru_RU-irina and
+// the other models this daemon has been pointed at). targetSampleRate is the
+// canonical maven wire rate (audio.PCM16kMono) that every downstream
+// consumer — playback, the voice wire, whisper on the way back in — expects.
+const (
+	piperSampleRate  = 22050
+	targetSampleRate = 16000
+)
+
+// resample22050To16000 converts raw 16-bit PCM from piperSampleRate to
+// targetSampleRate using linear interpolation.
 func resample22050To16000(input []byte) []byte {
 	if len(input) < 2 {
 		return nil
 	}
 
 	nSamples := len(input) / 2
-	outSamples := int(float64(nSamples) * 16000.0 / 22050.0)
+	outSamples := int(float64(nSamples) * float64(targetSampleRate) / float64(piperSampleRate))
 	output := make([]byte, outSamples*2)
 
-	ratio := 22050.0 / 16000.0
+	ratio := float64(piperSampleRate) / float64(targetSampleRate)
 
 	for i := 0; i < outSamples; i++ {
 		srcPos := float64(i) * ratio

cmd/mavwaked/main.go

@@ -5,12 +5,17 @@
 // is detected sends it as a PushToTalk frame to the voice server. The reply
 // audio is played back through aplay(1).
 //
-// No wake-word model yet (MVP uses voice-activity-only trigger). The
-// SurfaceVoice auth layer caps all commands at L0 (no destructive acts),
-// making accidental triggers safe by design. A proper wake-word engine
-// (openWakeWord / Silero VAD ONNX) is the planned upgrade — the VAD shape
-// (30ms frames, 16kHz PCM) matches silero-vad's input interface exactly, so
-// swapping energy-threshold for ONNX-inference is a local change in vad.go.
+// Voice activity is silero-vad when -vad-model points at the graph, and an
+// energy threshold when it does not. Silero declines noise the threshold
+// accepts: 0 frames against 68 to 99 on the four fixtures, measured in
+// docs/evals/2026-08-09-silero-vad.md. Note that the model window is 512
+// samples and the capture frame is 480, so silero.go re-chunks. This comment
+// used to say the two matched, which was true of silero v4.
+//
+// There is still no wake-word model, so anything spoken near the microphone
+// becomes a turn (V-487 stage two). The SurfaceVoice auth layer caps all
+// commands at L0 (no destructive acts), which is what makes an accidental
+// trigger safe rather than expensive.
 //
 // While a reply is playing the capture side is muted (half-duplex): without
 // it, Maven's own voice comes back in through the mic and she answers
@@ -73,6 +78,9 @@ func run(args []string) error {
 	bargeIn := flag.Bool("barge-in", false, "cut Maven off when he talks over her (needs a room-tuned -barge-in-rms)")
 	bargeRMS := flag.Int("barge-in-rms", defaultBargeRMS, "RMS x10000 a frame must clear to count as barge-in")
 	bargeFrames := flag.Int("barge-in-frames", defaultBargeFrames, "consecutive frames over -barge-in-rms before playback is cut")
+	vadModel := flag.String("vad-model", "", "silero-vad onnx file; empty runs the energy threshold instead")
+	vadThreshold := flag.Float64("vad-threshold", defaultSileroThreshold, "speech probability a frame must clear")
+	onnxLib := flag.String("onnx-lib", os.Getenv("MAVEN_ONNX_LIB"), "libonnxruntime.so, needed with -vad-model")
 	flag.CommandLine.Parse(args)
 
 	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM, syscall.SIGHUP)
@@ -82,8 +90,20 @@ func run(args []string) error {
 	vc := voice.Dial(*addr)
 	defer vc.Close()
 
-	// VAD engine.
+	// VAD engine. A model that will not load is logged and not fatal: the
+	// energy threshold is worse, and it is a great deal better than a
+	// listening client that refuses to start.
 	vad := NewVAD(*minRMS, *speechMs, *silenceMs, *maxMs)
+	if *vadModel != "" {
+		s, err := newSileroVAD(*vadModel, *onnxLib)
+		if err != nil {
+			log.Printf("mavwaked: silero unavailable, energy threshold unchanged: %v", err)
+		} else {
+			defer s.Close()
+			vad.UseSilero(s, *vadThreshold)
+			log.Printf("mavwaked: silero-vad from %s, threshold %.2f", *vadModel, *vadThreshold)
+		}
+	}
 
 	// Audio source.
 	var src io.ReadCloser

cmd/mavwaked/silero.go

@@ -0,0 +1,171 @@
+package main
+
+// silero-vad, the speech detector that replaces the energy threshold (V-487).
+//
+// Why an energy threshold is not a voice activity detector. It answers "is
+// this frame loud", and a fan, a door and a television are all loud. mavwaked
+// sends every utterance it accepts to speech-to-text and then to the daemon,
+// so a false trigger is a turn Maven takes on something nobody said to her.
+// Silero answers "is this frame speech", which is the question.
+//
+// It is 2.3MB of ONNX and runs on one CPU core in real time. That is not an
+// aside: this is the one model in the system that may never be offloaded or
+// gated on GPU admission, because a wake path that waits on a card is not a
+// wake path.
+//
+// Nil is a working value. Without -vad-model the daemon runs the energy VAD
+// exactly as it did before this file existed.
+
+import (
+	"fmt"
+	"sync"
+
+	ort "github.com/yalue/onnxruntime_go"
+)
+
+const (
+	// sileroWindow — samples per inference at 16kHz. The model is fixed at
+	// 512 and does not accept another size, which is why this file
+	// re-chunks rather than reusing the 480-sample capture frame. main.go
+	// used to claim the two matched; that was true of silero v4.
+	sileroWindow = 512
+
+	// sileroContext — samples of the previous window prepended to each
+	// inference, as the reference implementation does. Without it the first
+	// milliseconds of every window are judged with no history and speech
+	// onsets score low.
+	sileroContext = 64
+
+	// sileroState — the LSTM state carried between windows, [2][1][128].
+	sileroStateDim = 128
+
+	// defaultSileroThreshold — probability above which a window is speech.
+	// 0.5 is the reference default. Raising it costs speech onsets, which
+	// are the quietest part of an utterance.
+	defaultSileroThreshold = 0.5
+)
+
+// sileroVAD holds one ONNX session and the streaming state around it. It is
+// fed 30ms capture frames and answers per frame, buffering across calls
+// because 480 samples never line up with a 512-sample window.
+type sileroVAD struct {
+	mu      sync.Mutex
+	session *ort.DynamicAdvancedSession
+
+	pending []float32              // samples not yet part of a full window
+	context [sileroContext]float32 // tail of the previous window
+	state   []float32              // [2][1][128], carried between windows
+	last    float64                // most recent probability, held between windows
+	sr      []int64
+}
+
+// newSileroVAD loads the graph. The ONNX environment is initialised here when
+// nothing else has done it, because mavwaked has no embedder to do it first.
+func newSileroVAD(modelPath, libPath string) (*sileroVAD, error) {
+	if !ort.IsInitialized() {
+		if libPath != "" {
+			ort.SetSharedLibraryPath(libPath)
+		}
+		if err := ort.InitializeEnvironment(); err != nil {
+			return nil, fmt.Errorf("silero: onnx runtime: %w", err)
+		}
+	}
+	s, err := ort.NewDynamicAdvancedSession(modelPath,
+		[]string{"input", "state", "sr"}, []string{"output", "stateN"}, nil)
+	if err != nil {
+		return nil, fmt.Errorf("silero: load %s: %w", modelPath, err)
+	}
+	return &sileroVAD{
+		session: s,
+		state:   make([]float32, 2*sileroStateDim),
+		sr:      []int64{16000},
+	}, nil
+}
+
+// Speech reports whether the frame carries speech, and the probability behind
+// that answer. A frame that completes no window inherits the previous
+// probability, so the caller sees one answer per frame either way.
+func (s *sileroVAD) Speech(frame []int16, threshold float64) (bool, float64) {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	for _, v := range frame {
+		s.pending = append(s.pending, float32(v)/32768.0)
+	}
+	for len(s.pending) >= sileroWindow {
+		p, err := s.infer(s.pending[:sileroWindow])
+		if err != nil {
+			// A failed inference must not silence the microphone. Hold the
+			// last answer and let the next window try again.
+			break
+		}
+		s.last = p
+		s.pending = s.pending[sileroWindow:]
+	}
+	return s.last >= threshold, s.last
+}
+
+// infer runs one window and rolls the state and the context forward.
+func (s *sileroVAD) infer(window []float32) (float64, error) {
+	in := make([]float32, sileroContext+sileroWindow)
+	copy(in, s.context[:])
+	copy(in[sileroContext:], window)
+
+	inT, err := ort.NewTensor(ort.NewShape(1, int64(len(in))), in)
+	if err != nil {
+		return 0, err
+	}
+	defer inT.Destroy()
+	stT, err := ort.NewTensor(ort.NewShape(2, 1, sileroStateDim), s.state)
+	if err != nil {
+		return 0, err
+	}
+	defer stT.Destroy()
+	srT, err := ort.NewTensor(ort.NewShape(1), s.sr)
+	if err != nil {
+		return 0, err
+	}
+	defer srT.Destroy()
+
+	out, err := ort.NewEmptyTensor[float32](ort.NewShape(1, 1))
+	if err != nil {
+		return 0, err
+	}
+	defer out.Destroy()
+	next, err := ort.NewEmptyTensor[float32](ort.NewShape(2, 1, sileroStateDim))
+	if err != nil {
+		return 0, err
+	}
+	defer next.Destroy()
+
+	if err := s.session.Run(
+		[]ort.Value{inT, stT, srT},
+		[]ort.Value{out, next},
+	); err != nil {
+		return 0, err
+	}
+	copy(s.state, next.GetData())
+	copy(s.context[:], in[len(in)-sileroContext:])
+	return float64(out.GetData()[0]), nil
+}
+
+// Reset drops the streaming state. Called at every utterance boundary and
+// after barge-in, so echo-era history never scores the next sentence.
+func (s *sileroVAD) Reset() {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	s.pending = s.pending[:0]
+	s.context = [sileroContext]float32{}
+	for i := range s.state {
+		s.state[i] = 0
+	}
+	s.last = 0
+}
+
+// Close releases the session.
+func (s *sileroVAD) Close() error {
+	if s == nil || s.session == nil {
+		return nil
+	}
+	return s.session.Destroy()
+}

cmd/mavwaked/silero_test.go

@@ -0,0 +1,149 @@
+package main
+
+// What this measures. The energy threshold cannot tell a voice from a
+// television, and every utterance it accepts becomes a turn. So the test that
+// matters is not "does silero find speech" — it is "does it decline what the
+// energy threshold accepts".
+//
+// Speech is the four piper fixtures mavsttd already scores against. They are
+// synthesised, so nothing of the owner's voice is committed. Non-speech is
+// white noise at the same loudness, which is the cheapest thing that fools an
+// energy floor and the honest floor for this claim.
+//
+// Both halves skip without models/vad/silero_vad.onnx and MAVEN_ONNX_LIB,
+// like the TestONNX measurements in internal/router/eval.
+
+import (
+	"math"
+	"math/rand"
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+const wavHeader = 44 // 16kHz mono s16le, written by piper
+
+func loadSilero(t *testing.T) *sileroVAD {
+	t.Helper()
+	model := filepath.Join("..", "..", "models", "vad", "silero_vad.onnx")
+	lib := os.Getenv("MAVEN_ONNX_LIB")
+	if _, err := os.Stat(model); err != nil {
+		t.Skipf("missing %s: %v", model, err)
+	}
+	if lib == "" {
+		t.Skip("MAVEN_ONNX_LIB unset")
+	}
+	s, err := newSileroVAD(model, lib)
+	if err != nil {
+		t.Skipf("silero unavailable: %v", err)
+	}
+	return s
+}
+
+// feedAll runs a whole clip through a VAD and reports how many utterances it
+// produced and how many frames it called speech.
+func feedAll(v *VAD, pcm []int16) (utterances, speechFrames int) {
+	for i := 0; i+frameSamples <= len(pcm); i += frameSamples {
+		frame := pcm[i : i+frameSamples]
+		utt, state := v.Feed(frame)
+		if state == StateSpeech {
+			speechFrames++
+		}
+		if utt.Bytes != nil {
+			utterances++
+		}
+	}
+	return utterances, speechFrames
+}
+
+func readFixture(t *testing.T, name string) []int16 {
+	t.Helper()
+	raw, err := os.ReadFile(filepath.Join("..", "mavsttd", "testdata", name))
+	if err != nil {
+		t.Skipf("missing fixture %s: %v", name, err)
+	}
+	if len(raw) <= wavHeader {
+		t.Fatalf("%s: %d bytes, no audio", name, len(raw))
+	}
+	return PCMToI16(raw[wavHeader:])
+}
+
+// noise returns white noise scaled to the same RMS as ref. Same loudness,
+// nothing said.
+func noise(ref []int16, seed int64) []int16 {
+	target := frameRMS(ref)
+	r := rand.New(rand.NewSource(seed))
+	out := make([]int16, len(ref))
+	for i := range out {
+		out[i] = int16(r.NormFloat64() * target * 32768.0)
+	}
+	return out
+}
+
+func TestSileroHearsSpeechAndDeclinesNoise(t *testing.T) {
+	s := loadSilero(t)
+	defer s.Close()
+
+	for _, name := range []string{"ru_fact.wav", "ru_query.wav", "ru_reminder.wav", "en_act.wav"} {
+		pcm := readFixture(t, name)
+
+		v := NewVAD(0, 0, 0, 0)
+		v.UseSilero(s, defaultSileroThreshold)
+		_, spoke := feedAll(v, pcm)
+		if spoke == 0 {
+			t.Errorf("%s: silero heard no speech in a spoken clip", name)
+		}
+
+		s.Reset()
+		v2 := NewVAD(0, 0, 0, 0)
+		v2.UseSilero(s, defaultSileroThreshold)
+		_, heard := feedAll(v2, noise(pcm, 7))
+
+		energy := NewVAD(0, 0, 0, 0)
+		_, energyHeard := feedAll(energy, noise(pcm, 7))
+
+		t.Logf("%s: speech frames — silero on speech %d, silero on noise %d, energy on noise %d",
+			name, spoke, heard, energyHeard)
+		if heard >= energyHeard {
+			t.Errorf("%s: silero called %d noise frames speech, energy called %d — no improvement",
+				name, heard, energyHeard)
+		}
+		s.Reset()
+	}
+}
+
+// BenchmarkSileroFrame answers the only performance question that matters
+// here: one 30ms frame must cost far less than 30ms on one core, or the
+// detector cannot run always-on beside everything else on that machine.
+func BenchmarkSileroFrame(b *testing.B) {
+	s := loadSilero(&testing.T{})
+	if s == nil {
+		b.Skip("silero unavailable")
+	}
+	defer s.Close()
+	frame := make([]int16, frameSamples)
+	for i := range frame {
+		frame[i] = int16(i%400 - 200)
+	}
+	for i := 0; i < b.N; i++ {
+		s.Speech(frame, defaultSileroThreshold)
+	}
+}
+
+// TestSileroRechunksAcrossFrames pins the reason this file exists. The capture
+// frame is 480 samples and the model window is 512, so a detector that ran one
+// inference per frame would be feeding the model a shape it does not accept.
+func TestSileroRechunksAcrossFrames(t *testing.T) {
+	s := loadSilero(t)
+	defer s.Close()
+
+	silence := make([]int16, frameSamples)
+	for i := 0; i < 20; i++ {
+		if _, p := s.Speech(silence, defaultSileroThreshold); math.IsNaN(p) {
+			t.Fatalf("frame %d: probability is NaN", i)
+		}
+	}
+	if len(s.pending) >= sileroWindow {
+		t.Errorf("pending grew to %d samples, so windows are not being consumed", len(s.pending))
+	}
+}

cmd/mavwaked/vad.go

@@ -68,6 +68,37 @@ type VAD struct {
 	// follows the room's ambient level. Initialised to minRMS; updated
 	// on each silence frame.
 	floorRMS float64
+
+	// speech is silero-vad, or nil. When it is set the energy floor decides
+	// nothing: the question becomes "is this speech" rather than "is this
+	// loud", and the noise floor is not even tracked. Everything after that
+	// answer — the speech hold, the silence hold, the length cap, the
+	// buffer — is the same state machine either way, which is why the
+	// detector goes here and not around this type.
+	speech    *sileroVAD
+	speechMin float64
+}
+
+// UseSilero swaps the energy threshold for the model. Passing nil is a
+// no-op, so a caller that could not load the graph keeps a working VAD.
+func (v *VAD) UseSilero(s *sileroVAD, threshold float64) {
+	if s == nil {
+		return
+	}
+	if threshold <= 0 {
+		threshold = defaultSileroThreshold
+	}
+	v.speech = s
+	v.speechMin = threshold
+}
+
+// isSpeech answers the one question the state machine asks of a frame.
+func (v *VAD) isSpeech(frame []int16, rms float64) bool {
+	if v.speech != nil {
+		ok, _ := v.speech.Speech(frame, v.speechMin)
+		return ok
+	}
+	return rms >= v.floorRMS
 }
 
 // NewVAD creates a VAD with the given thresholds. Zero values use defaults.
@@ -110,7 +141,7 @@ func (v *VAD) State() SpeechState { return v.state }
 // should send the audio to the voice server before feeding more frames.
 func (v *VAD) Feed(frame []int16) (_ audio.Audio, state SpeechState) {
 	rms := frameRMS(frame)
-	isSpeech := rms >= v.floorRMS
+	isSpeech := v.isSpeech(frame, rms)
 
 	switch v.state {
 	case StateSilence:
@@ -175,6 +206,9 @@ func (v *VAD) Feed(frame []int16) (_ audio.Audio, state SpeechState) {
 func (v *VAD) Reset() { v.reset() }
 
 func (v *VAD) reset() {
+	if v.speech != nil {
+		v.speech.Reset()
+	}
 	v.state = StateSilence
 	v.speechFrames = 0
 	v.silenceFrames = 0

cmd/mavweb/chat.go

@@ -0,0 +1,195 @@
+package main
+
+import (
+	_ "embed"
+	"errors"
+	"log"
+	"net/http"
+	"net/url"
+	"strconv"
+	"strings"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/router"
+	"github.com/kami/maven/internal/webauthn"
+)
+
+//go:embed chat.html
+var chatPageHTML string
+
+// chatTmpl — plain text conversation interface. No JS: form POSTs to /api/chat
+// and the handler redirects back to /chat with the response.
+var chatTmpl = parsePage("chat", chatPageHTML, nil)
+
+// chatMsg — one message in the conversation history.
+type chatMsg struct {
+	Role string // "user" | "assistant"
+	Text string
+	// Source — the query source that claimed the turn, shown as a badge beside
+	// the reply. Empty for a turn no source claimed (V-539).
+	Source string
+	// TraceID anchors the correction gesture (V-630). Non-zero ⇒ the turn was
+	// persisted and can be corrected in one click. 0 ⇒ no correction is offered,
+	// which is honest: a box with no database has no turn to correct.
+	TraceID int64
+	// Corrected — the owner already corrected this turn, so the page says thank
+	// you instead of offering the buttons again.
+	Corrected string
+}
+
+// correctionTargets — the seven public intents, in the order the buttons are
+// shown. Read from internal/router rather than typed out, so a new intent cannot
+// exist without a way to correct a turn into it.
+var correctionTargets = []router.Intent{
+	router.IntentFact, router.IntentNote, router.IntentReminder,
+	router.IntentQuery, router.IntentAct, router.IntentChat, router.IntentSystem,
+}
+
+// handleChatPage renders the chat conversation page.
+func handleChatPage(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "chat") {
+		return
+	}
+	msgs := []chatMsg{}
+	// Read user message + reply from query params (set by /api/chat redirect).
+	if q := r.URL.Query().Get("q"); q != "" {
+		msgs = append(msgs, chatMsg{Role: "user", Text: q})
+	}
+	if reply := r.URL.Query().Get("r"); reply != "" {
+		id, _ := strconv.ParseInt(r.URL.Query().Get("t"), 10, 64)
+		msgs = append(msgs, chatMsg{
+			Role: "assistant", Text: reply, Source: r.URL.Query().Get("s"),
+			TraceID: id, Corrected: r.URL.Query().Get("c"),
+		})
+	}
+	// UserText rides beside the messages so the correction form can hand the
+	// conversation back on the redirect: this page has no session and no JS, so
+	// what is on screen is what the query params carry.
+	renderPage(w, chatTmpl, struct {
+		Error    string
+		Messages []chatMsg
+		Targets  []router.Intent
+		UserText string
+	}{Messages: msgs, Targets: correctionTargets, UserText: r.URL.Query().Get("q")})
+}
+
+// handleChatAPI processes a chat message POST and redirects back to /chat.
+//
+// State-changing, and the widest surface on this server: the text reaches the
+// router, the LLM, and through mavend's applyAction the whole action path
+// including `act` — so it is gated on the same step-up as POST /tools and
+// POST /api/revert (Vikunja #317). With WebAuthn unconfigured the gate is
+// fail-open exactly like the others (see stepUpOK); with -require-stepup it
+// denies, which is the point of that flag.
+func handleChatAPI(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, session *webauthn.PasskeySession, requireStepUp bool) {
+	if r.Method != http.MethodPost {
+		http.Error(w, "POST only", http.StatusMethodNotAllowed)
+		return
+	}
+	if !requireCore(w, core, "chat") {
+		return
+	}
+	if !stepUpGate(w, session, requireStepUp) {
+		return
+	}
+	text := strings.TrimSpace(r.FormValue("text"))
+	if text == "" {
+		http.Redirect(w, r, "/chat", http.StatusSeeOther)
+		return
+	}
+	// One conversation id for the whole web chat, and a different one from
+	// telegram or the mic. A parked question belongs to the reach that was
+	// asked; before this, a clarify nobody answered on the web ate the next
+	// utterance spoken at the mic (Vikunja #466). This server has no
+	// per-browser session, so every browser tab is the same conversation —
+	// which is right for a single-owner box.
+	reply, err := core.Chat(r.Context(), "web", text)
+	if err != nil {
+		log.Printf("chat api: %v", err)
+		http.Redirect(w, r, "/chat", http.StatusSeeOther)
+		return
+	}
+	// The claiming query source rides back on the redirect so the page can show
+	// it. Empty for a turn no source claimed, which is most of them.
+	dest := "/chat?q=" + url.QueryEscape(text) + "&r=" + url.QueryEscape(reply.Reply)
+	if reply.Source != "" {
+		dest += "&s=" + url.QueryEscape(reply.Source)
+	}
+	// The trace id rides along so the reply can carry a correction gesture
+	// (V-630). Absent when nothing persisted, and the page then offers none.
+	if reply.TraceID != 0 {
+		dest += "&t=" + strconv.FormatInt(reply.TraceID, 10)
+	}
+	http.Redirect(w, r, dest, http.StatusSeeOther)
+}
+
+// handleCorrectAPI records that the last turn was routed wrongly (V-630).
+//
+// A correction is the only supervised signal this box gets, and everything else
+// in the trace accumulates on its own. So the gesture has to cost nothing: one
+// POST from the reply he is already looking at, carrying the trace id and
+// optionally the intent it should have been. An unstated target is accepted,
+// because a turn marked wrong with no target is still a usable negative.
+//
+// Step-up gated like POST /api/chat, and that costs the gesture nothing: he
+// tapped to send the turn he is now correcting, so the session is already up.
+// It is gated because trace ids are sequential integers and this writes the one
+// table the routing heads (V-546) will be fitted on. A caller who can guess an
+// id could otherwise mislabel turns he never corrected.
+func handleCorrectAPI(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, session *webauthn.PasskeySession, requireStepUp bool) {
+	if r.Method != http.MethodPost {
+		http.Error(w, "POST only", http.StatusMethodNotAllowed)
+		return
+	}
+	if !requireCore(w, core, "correct") {
+		return
+	}
+	if !stepUpGate(w, session, requireStepUp) {
+		return
+	}
+	id, err := strconv.ParseInt(strings.TrimSpace(r.FormValue("trace_id")), 10, 64)
+	if err != nil || id <= 0 {
+		http.Error(w, "trace_id required", http.StatusBadRequest)
+		return
+	}
+	shouldBe := strings.TrimSpace(r.FormValue("should_be"))
+	// Only one of the seven, or nothing. Free text here would put an unroutable
+	// label in the one table V-632 fits prototypes from.
+	if shouldBe != "" && !isCorrectionTarget(shouldBe) {
+		http.Error(w, "should_be must be one of the seven intents", http.StatusBadRequest)
+		return
+	}
+	if err := core.CorrectTurn(r.Context(), id, shouldBe); err != nil {
+		log.Printf("correct turn %d: %v", id, err)
+		// A turn past the retention bound is gone, and saying so is different
+		// from saying the write broke.
+		if errors.Is(err, ipc.ErrNoSuchTrace) {
+			http.Error(w, "that turn is no longer stored", http.StatusNotFound)
+			return
+		}
+		http.Error(w, "correction failed", http.StatusBadGateway)
+		return
+	}
+	stamp := shouldBe
+	if stamp == "" {
+		stamp = "wrong"
+	}
+	// Back to the conversation he was in, with the turn still on screen. The
+	// query params carry it, so the correction is preserved by re-sending them.
+	dest := "/chat?q=" + url.QueryEscape(r.FormValue("q")) +
+		"&r=" + url.QueryEscape(r.FormValue("rep")) + "&c=" + url.QueryEscape(stamp)
+	if s := r.FormValue("s"); s != "" {
+		dest += "&s=" + url.QueryEscape(s)
+	}
+	http.Redirect(w, r, dest, http.StatusSeeOther)
+}
+
+// isCorrectionTarget — one of the seven, and nothing else.
+func isCorrectionTarget(s string) bool {
+	for _, t := range correctionTargets {
+		if string(t) == s {
+			return true
+		}
+	}
+	return false
+}

cmd/mavweb/chat.html

@@ -5,6 +5,21 @@
 <div class="scroll chat-scroll" id=chatHistory>
 {{range .Messages}}
   <div class="chat-msg {{.Role}}"><strong>{{if eq .Role "user"}}you{{else}}maven{{end}}:</strong> {{.Text}}{{if .Source}} <span class="badge badge-accent" title="the query source that claimed this turn">{{.Source}}</span>{{end}}</div>
+  {{if and (eq .Role "assistant") .TraceID}}
+  {{if .Corrected}}
+  <div class=chat-correct><span class="badge badge-ok" title="the label is kept; the transcript still expires in 14 days">corrected: {{.Corrected}}</span></div>
+  {{else}}
+  <form method=post action=/api/correct class=chat-correct>
+    <input type=hidden name=trace_id value="{{.TraceID}}">
+    <input type=hidden name=q value="{{$.UserText}}">
+    <input type=hidden name=rep value="{{.Text}}">
+    <input type=hidden name=s value="{{.Source}}">
+    <button class="btn btn-sm" title="wrong, and I am not saying what it was">wrong</button>
+    <span class=chat-correct-label>should have been:</span>
+    {{range $.Targets}}<button class="btn btn-sm btn-muted" name=should_be value="{{.}}">{{.}}</button>{{end}}
+  </form>
+  {{end}}
+  {{end}}
 {{else}}
   <div class=empty>
     <svg class=icon width="20" height="20"><use href="/ethos-icons.svg#i-message"/></svg>

cmd/mavweb/correct_test.go

@@ -0,0 +1,160 @@
+package main
+
+import (
+	"context"
+	"errors"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"strings"
+	"testing"
+
+	"github.com/kami/maven/internal/ipc"
+)
+
+// correctCore records the correction the handler sends.
+type correctCore struct {
+	ipc.UnimplementedCoreAPI
+	traceID  int64
+	shouldBe string
+	called   bool
+	err      error
+}
+
+func (c *correctCore) CorrectTurn(_ context.Context, traceID int64, shouldBe string) error {
+	c.called, c.traceID, c.shouldBe = true, traceID, shouldBe
+	return c.err
+}
+
+func postCorrect(form url.Values) *http.Request {
+	req := httptest.NewRequest(http.MethodPost, "/api/correct", strings.NewReader(form.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	return req
+}
+
+// The full gesture: wrong, and it should have been a fact.
+func TestCorrectAPIWithTarget(t *testing.T) {
+	core := &correctCore{}
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, postCorrect(url.Values{
+		"trace_id": {"42"}, "should_be": {"fact"}, "q": {"поужинал"}, "rep": {"поняла"},
+	}), core, stepUpSession(), false)
+
+	if rr.Code != http.StatusSeeOther {
+		t.Fatalf("status %d, want 303; body=%s", rr.Code, rr.Body.String())
+	}
+	if core.traceID != 42 || core.shouldBe != "fact" {
+		t.Errorf("corrected trace %d to %q", core.traceID, core.shouldBe)
+	}
+	// The turn stays on screen, and the page says it was corrected.
+	loc := rr.Header().Get("Location")
+	if !strings.Contains(loc, "c=fact") || !strings.Contains(loc, "q=") {
+		t.Errorf("redirect %q loses the turn or the correction", loc)
+	}
+}
+
+// The cheap half. A turn marked wrong with no target is still a usable negative,
+// and it must not cost more to give than the full answer.
+func TestCorrectAPIWithNoTarget(t *testing.T) {
+	core := &correctCore{}
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, postCorrect(url.Values{"trace_id": {"7"}}), core, stepUpSession(), false)
+
+	if rr.Code != http.StatusSeeOther {
+		t.Fatalf("status %d, want 303", rr.Code)
+	}
+	if !core.called || core.shouldBe != "" {
+		t.Errorf("called=%v shouldBe=%q, want an untargeted negative recorded", core.called, core.shouldBe)
+	}
+	if !strings.Contains(rr.Header().Get("Location"), "c=wrong") {
+		t.Errorf("redirect %q does not say the turn was marked wrong", rr.Header().Get("Location"))
+	}
+}
+
+// Free text here would put an unroutable label in the one table V-632 fits
+// prototypes from.
+func TestCorrectAPIRejectsUnknownTarget(t *testing.T) {
+	core := &correctCore{}
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, postCorrect(url.Values{"trace_id": {"7"}, "should_be": {"погода"}}), core, stepUpSession(), false)
+
+	if rr.Code != http.StatusBadRequest {
+		t.Fatalf("status %d, want 400", rr.Code)
+	}
+	if core.called {
+		t.Error("wrote a label for a target that is not one of the seven")
+	}
+}
+
+func TestCorrectAPINeedsTraceID(t *testing.T) {
+	for _, form := range []url.Values{{}, {"trace_id": {"0"}}, {"trace_id": {"nope"}}} {
+		core := &correctCore{}
+		rr := httptest.NewRecorder()
+		handleCorrectAPI(rr, postCorrect(form), core, stepUpSession(), false)
+		if rr.Code != http.StatusBadRequest {
+			t.Errorf("form %v: status %d, want 400", form, rr.Code)
+		}
+		if core.called {
+			t.Errorf("form %v: reached the core", form)
+		}
+	}
+}
+
+// A write that broke is not a turn that expired, and the two must not read the
+// same to the owner deciding whether to correct again.
+func TestCorrectAPIReportsFailure(t *testing.T) {
+	core := &correctCore{err: errors.New("disk is full")}
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, postCorrect(url.Values{"trace_id": {"9"}, "should_be": {"note"}}), core, stepUpSession(), false)
+	if rr.Code != http.StatusBadGateway {
+		t.Fatalf("status %d, want 502", rr.Code)
+	}
+}
+
+// A trace past the retention bound is gone, and the surface says that.
+func TestCorrectAPIExpiredTurn(t *testing.T) {
+	core := &correctCore{err: ipc.ErrNoSuchTrace}
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, postCorrect(url.Values{"trace_id": {"9"}, "should_be": {"note"}}), core, stepUpSession(), false)
+	if rr.Code != http.StatusNotFound {
+		t.Fatalf("status %d, want 404", rr.Code)
+	}
+}
+
+func TestCorrectAPIPostOnly(t *testing.T) {
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, httptest.NewRequest(http.MethodGet, "/api/correct", nil), &correctCore{}, stepUpSession(), false)
+	if rr.Code != http.StatusMethodNotAllowed {
+		t.Fatalf("status %d, want 405", rr.Code)
+	}
+}
+
+// Every one of the seven intents has a button, so a new intent cannot exist with
+// no way to correct a turn into it.
+func TestCorrectionTargetsAreTheSeven(t *testing.T) {
+	if len(correctionTargets) != 7 {
+		t.Fatalf("%d targets, want the seven public intents", len(correctionTargets))
+	}
+	for _, want := range []string{"fact", "note", "reminder", "query", "act", "chat", "system"} {
+		if !isCorrectionTarget(want) {
+			t.Errorf("%s is not offered", want)
+		}
+	}
+	if isCorrectionTarget("") {
+		t.Error("empty is not a target: it is the absence of one, handled separately")
+	}
+}
+
+// Trace ids are sequential, so a caller who cannot assert step-up must not be
+// able to label a turn the owner never corrected.
+func TestCorrectAPINeedsStepUp(t *testing.T) {
+	core := &correctCore{}
+	rr := httptest.NewRecorder()
+	handleCorrectAPI(rr, postCorrect(url.Values{"trace_id": {"9"}, "should_be": {"note"}}), core, nil, true)
+	if rr.Code != http.StatusForbidden {
+		t.Fatalf("status %d, want 403", rr.Code)
+	}
+	if core.called {
+		t.Error("wrote a label with no step-up")
+	}
+}

cmd/mavweb/ecosystem.go

@@ -117,8 +117,5 @@ func handleEcosystem(w http.ResponseWriter, r *http.Request, urls ecoURLs, core
 		d.Calls.Rows = rows
 	}
 
-	w.Header().Set("Content-Type", "text/html; charset=utf-8")
-	if err := ecosystemTmpl.Execute(w, d); err != nil {
-		log.Printf("ecosystem render: %v", err)
-	}
+	renderPage(w, ecosystemTmpl, d)
 }

cmd/mavweb/events.html

@@ -14,8 +14,13 @@ memory only, so a restart empties this.</div>
 <div class=scroll><table class=mono>
 <tr><th>noticed<th>happened<th>source<th>kind<th>pri<th>what<th>detail</tr>
 {{range .Events}}<tr>
-<td>{{.NoticedAt.Format "02.01 15:04:05"}}</td>
-<td class=gray>{{.OccurredAt.Format "02.01 15:04:05"}}</td>
+<!-- Both columns in his clock (V-469 on /reminders, same rule here). NoticedAt
+     is the bus's local instant, OccurredAt is whatever zone the source used —
+     the store hands back UTC and internal/rss parses a pubDate to UTC — so
+     rendering them raw put two zones side by side in the same row and made a
+     feed item look hours older than it was. -->
+<td>{{.NoticedAt.Local.Format "02.01 15:04:05"}}</td>
+<td class=gray>{{.OccurredAt.Local.Format "02.01 15:04:05"}}</td>
 <td class=gray>{{.Source}}</td>
 <td class=gray>{{.Kind}}</td>
 <td class=gray>{{.Priority}}</td>

cmd/mavweb/events_test.go

@@ -46,7 +46,8 @@ func TestEventsPageRendersTheJournal(t *testing.T) {
 		t.Fatalf("status = %d, want 200", w.Code)
 	}
 	body := w.Body.String()
-	for _, want := range []string{"rss:tech", "Вышло ядро 6.19", "ambient:notif", "10:00-11:00 планёрка", "01.08 10:00:00"} {
+	occurred := time.Date(2026, 8, 1, 10, 0, 0, 0, time.UTC).Local().Format("02.01 15:04:05")
+	for _, want := range []string{"rss:tech", "Вышло ядро 6.19", "ambient:notif", "10:00-11:00 планёрка", occurred} {
 		if !strings.Contains(body, want) {
 			t.Errorf("page does not mention %q", want)
 		}
@@ -89,6 +90,38 @@ func TestEventsPageWithoutCore(t *testing.T) {
 	}
 }
 
+// awayFromLocal returns a zone three hours off whatever this machine runs in,
+// so a test can tell "rendered in his clock" apart from "rendered in whatever
+// zone the value arrived in" without depending on TZ.
+func awayFromLocal() *time.Location {
+	_, off := time.Now().Zone()
+	return time.FixedZone("away", off+3*60*60)
+}
+
+func TestEventsPageRendersBothTimesInLocalZone(t *testing.T) {
+	// OccurredAt carries the source's zone — the store hands back UTC and
+	// internal/rss parses a pubDate to UTC — while NoticedAt is the bus's local
+	// instant. Rendered raw, the two columns of one row were in two zones and a
+	// feed item read hours older than it was.
+	away := awayFromLocal()
+	occurred := time.Date(2026, 8, 1, 7, 15, 0, 0, time.UTC).In(away)
+	noticed := occurred.Add(2 * time.Minute)
+	core := &eventsCore{events: []ipc.IntakeEvent{{
+		Source: "rss:tech", Kind: "note", Title: "Вышло ядро 6.19", Priority: "low",
+		OccurredAt: occurred, NoticedAt: noticed,
+	}}}
+	body := getEvents(t, core).Body.String()
+	const layout = "02.01 15:04:05"
+	for _, ts := range []time.Time{occurred, noticed} {
+		if !strings.Contains(body, ts.Local().Format(layout)) {
+			t.Errorf("page does not render %s in his clock (%s)", ts, ts.Local().Format(layout))
+		}
+		if strings.Contains(body, ts.In(away).Format(layout)) {
+			t.Errorf("page rendered %s in the source's zone", ts)
+		}
+	}
+}
+
 func TestEventsPageEscapesIntakeText(t *testing.T) {
 	// Titles come from outside — a feed headline, a notification. They are shown
 	// on a page and must never be able to inject markup into it.

cmd/mavweb/facts.go

@@ -0,0 +1,96 @@
+package main
+
+import (
+	"encoding/json"
+	"errors"
+	"log"
+	"net/http"
+	"strings"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/webauthn"
+)
+
+// The two fact-writing API routes: POST /api/signal appends a presence
+// observation, POST /api/revert voids the latest fact for a key. Neither
+// renders a page.
+
+// presenceSignals — the only fact keys /api/signal may write. mavweb is a
+// network-facing surface inside wg; an allowlist keeps a compromised caller
+// boxed to forging weak presence signals (reachability, multi-source, never
+// truth) — it can't write arbitrary facts. ponytail: floor auth (wg-only); a
+// per-signal token belongs here if the tunnel ever hosts untrusted devices.
+var presenceSignals = map[string]string{
+	"desk_active":    "infer:hyprland",
+	"page_heartbeat": "infer:heartbeat",
+	"wg_handshake":   "infer:wg",
+}
+
+// handleSignal ingests one presence signal and writes a fresh fact through
+// CoreAPI. The fact's timestamp (now) is all the presence scorer reads; value
+// is a marker. Only allowlisted keys are accepted (see presenceSignals).
+func handleSignal(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if r.Method != http.MethodPost {
+		http.Error(w, "POST only", http.StatusMethodNotAllowed)
+		return
+	}
+	if !requireCore(w, core, "presence ingest") {
+		return
+	}
+	key := r.URL.Query().Get("key")
+	source, ok := presenceSignals[key]
+	if !ok {
+		http.Error(w, "unknown signal key", http.StatusBadRequest)
+		return
+	}
+	// kind=env: an observation about the device/surface, NOT a self-fact — a
+	// passive signal never writes truth about you (spec), it only feeds
+	// presence. confidence 1.0: the reading ("input happened") is certain;
+	// presence applies its own per-signal weight/decay on top.
+	if _, err := core.WriteFact(r.Context(), ipc.WriteFactReq{
+		Ts:         time.Now(),
+		Kind:       "env",
+		Key:        key,
+		Value:      `"active"`,
+		Source:     source,
+		Confidence: 1.0,
+	}); err != nil {
+		log.Printf("signal %s: %v", key, err)
+		http.Error(w, "write failed", http.StatusBadGateway)
+		return
+	}
+	w.WriteHeader(http.StatusNoContent)
+}
+
+func handleRevert(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, session *webauthn.PasskeySession, requireStepUp bool) {
+	if r.Method != http.MethodPost {
+		http.Error(w, "POST only", http.StatusMethodNotAllowed)
+		return
+	}
+	if !requireCore(w, core, "revert") {
+		return
+	}
+	if !stepUpGate(w, session, requireStepUp) {
+		return
+	}
+
+	key := strings.TrimSpace(r.FormValue("key"))
+	if key == "" {
+		http.Error(w, "key required", http.StatusBadRequest)
+		return
+	}
+	newID, err := core.RevertFact(r.Context(), key)
+	if err != nil {
+		log.Printf("revert %q: %v", key, err)
+		if errors.Is(err, ipc.ErrNoFact) {
+			http.Error(w, "no fact to revert", http.StatusNotFound)
+			return
+		}
+		http.Error(w, "revert failed", http.StatusBadGateway)
+		return
+	}
+	log.Printf("reverted fact for key=%s, new_id=%d", key, newID)
+	w.Header().Set("Content-Type", "application/json")
+	json.NewEncoder(w).Encode(map[string]any{"reverted": true, "new_id": newID})
+}

cmd/mavweb/handlers_test.go

@@ -65,6 +65,7 @@ type fakeCore struct {
 	// for handleTrace tests
 	tickTrace ipc.TickTrace
 	traceErr  error
+	turns     []ipc.TurnDecision
 
 	// for handleChatAPI tests
 	chatText   string
@@ -173,6 +174,10 @@ func (f *fakeCore) RevertFact(_ context.Context, key string) (int64, error) {
 	return f.revertNewID, nil
 }
 
+func (f *fakeCore) TurnDecisions(_ context.Context, _ int) ([]ipc.TurnDecision, error) {
+	return f.turns, nil
+}
+
 func (f *fakeCore) TickTrace(_ context.Context) (ipc.TickTrace, error) {
 	if f.traceErr != nil {
 		return ipc.TickTrace{}, f.traceErr
@@ -825,6 +830,41 @@ func TestHandleTrace(t *testing.T) {
 			t.Error("rendered 'nothing fired' but a winner was set")
 		}
 	})
+
+	// The turn arbitration shares this page (V-564). A reader must see the
+	// winner, a loser and the claimants that were never asked, because the last
+	// of those is what the hardcoded ordering hides.
+	t.Run("renders the turn decision record", func(t *testing.T) {
+		core := &fakeCore{turns: []ipc.TurnDecision{{
+			Ts:        time.Date(2025, 6, 1, 12, 0, 0, 0, time.UTC),
+			Utterance: "какая погода в риме",
+			Winner:    "query:weather",
+			Claims: []ipc.TurnClaim{
+				{Stage: "query", Claimant: "weather", Intent: "query", Outcome: "won"},
+				{Stage: "query", Claimant: "calendar", Outcome: "declined", Reason: "no answer"},
+				{Stage: "query", Claimant: "kiwix", Outcome: "never_asked"},
+			},
+		}}}
+		rr := httptest.NewRecorder()
+		handleTrace(rr, httptest.NewRequest(http.MethodGet, "/trace", nil), core)
+		body := rr.Body.String()
+		for _, want := range []string{"какая погода в риме", "query:weather", "calendar", "kiwix", "never_asked"} {
+			if !strings.Contains(body, want) {
+				t.Errorf("rendered page is missing %q", want)
+			}
+		}
+	})
+
+	t.Run("no turns renders the empty note, not an error", func(t *testing.T) {
+		rr := httptest.NewRecorder()
+		handleTrace(rr, httptest.NewRequest(http.MethodGet, "/trace", nil), &fakeCore{})
+		if rr.Code != http.StatusOK {
+			t.Fatalf("status = %d, want 200", rr.Code)
+		}
+		if !strings.Contains(rr.Body.String(), "no turn has run") {
+			t.Error("empty ring did not render its note")
+		}
+	})
 }
 
 // --- handleRevert ---

cmd/mavweb/models.go

@@ -2,8 +2,8 @@ package main
 
 import (
 	"context"
+	_ "embed"
 	"errors"
-	"html/template"
 	"log"
 	"net/http"
 	"strconv"
@@ -31,43 +31,10 @@ type modelController interface {
 	SwapModel(ctx context.Context, req ipc.SwapModelReq) (ipc.SwapModelResp, error)
 }
 
-var modelsTmpl = template.Must(template.New("models").Funcs(shellFuncs()).Parse(shellHTML + modelsHTML))
+//go:embed models.html
+var modelsHTML string
 
-const modelsHTML = `{{template "shellTop" "models"}}
-<h1>Resident model</h1>
-<p class=hint>swapping requires step-up — <a href=/auth/passkey>assert a passkey</a> first. The old model is unloaded before the new one is loaded (one model fits the iGPU at a time), so turns during the load are refused and fall back to the classifier.</p>
-<p class=hint>a swap is not remembered. Nothing writes it down, so the next restart of the daemon — including the one <code>mavupdate</code> does — comes back on <code>phraser.model_path</code> from the config. Make it stick by editing that.</p>
-{{if .Msg}}<div class="msg msg-ok">{{.Msg}}</div>{{end}}
-{{if .Err}}<div class="msg msg-err">{{.Err}}</div>{{end}}
-{{if .Off}}
-<section class=card>
-<h2 class=card-title>swap not configured</h2>
-<p class=hint>this core has no <code>phraser.swap_models</code> allowlist, so there is nothing to swap to. Add the gguf paths you allow to <code>deploy/mavend.json</code> and restart once.</p>
-</section>
-{{else}}
-<section class=card>
-<h2 class=card-title>loaded now</h2>
-<div class=scroll><table>
-<tr><th>model</th><td><code>{{.Status.Model}}</code></td></tr>
-<tr><th>file</th><td><code>{{.Status.ModelPath}}</code></td></tr>
-<tr><th>server</th><td><code>{{.Status.BaseURL}}</code></td></tr>
-<tr><th>n_ctx</th><td>{{.Status.NCtx}}</td></tr>
-<tr><th>n_gpu_layers</th><td>{{.Status.NGpuLayers}}</td></tr>
-</table></div>
-<p class=hint>the model name is what llama-server reports for itself, not what the config says it should be.</p>
-</section>
-<section class=card>
-<h2 class=card-title>allowed models <span class=badge>{{len .Status.Swappable}}</span></h2>
-{{if .Status.Swappable}}<div class=scroll><table><tr><th>file</th><th></th></tr>
-{{range .Status.Swappable}}<tr><td><code>{{.}}</code></td>
-<td><form method=post action=/models class=inline-form>
-<input type=hidden name=model_path value="{{.}}">
-<button class=btn>load this one</button></form></td></tr>{{end}}
-</table></div>
-{{else}}<div class=empty><div>no models allowlisted</div></div>{{end}}
-</section>
-{{end}}
-{{template "shellBottom"}}`
+var modelsTmpl = parsePage("models", modelsHTML, nil)
 
 type modelsPage struct {
 	Msg    string
@@ -154,8 +121,5 @@ func handleModels(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, swap
 		}
 	}
 	page.Status = st
-	w.Header().Set("Content-Type", "text/html; charset=utf-8")
-	if err := modelsTmpl.Execute(w, page); err != nil {
-		log.Printf("models render: %v", err)
-	}
+	renderPage(w, modelsTmpl, page)
 }

cmd/mavweb/models.html

@@ -0,0 +1,35 @@
+{{template "shellTop" "models"}}
+<h1>Resident model</h1>
+<p class=hint>swapping requires step-up — <a href=/auth/passkey>assert a passkey</a> first. The old model is unloaded before the new one is loaded (one model fits the iGPU at a time), so turns during the load are refused and fall back to the classifier.</p>
+<p class=hint>a swap is not remembered. Nothing writes it down, so the next restart of the daemon — including the one <code>mavupdate</code> does — comes back on <code>phraser.model_path</code> from the config. Make it stick by editing that.</p>
+{{if .Msg}}<div class="msg msg-ok">{{.Msg}}</div>{{end}}
+{{if .Err}}<div class="msg msg-err">{{.Err}}</div>{{end}}
+{{if .Off}}
+<section class=card>
+<h2 class=card-title>swap not configured</h2>
+<p class=hint>this core has no <code>phraser.swap_models</code> allowlist, so there is nothing to swap to. Add the gguf paths you allow to <code>deploy/mavend.json</code> and restart once.</p>
+</section>
+{{else}}
+<section class=card>
+<h2 class=card-title>loaded now</h2>
+<div class=scroll><table>
+<tr><th>model</th><td><code>{{.Status.Model}}</code></td></tr>
+<tr><th>file</th><td><code>{{.Status.ModelPath}}</code></td></tr>
+<tr><th>server</th><td><code>{{.Status.BaseURL}}</code></td></tr>
+<tr><th>n_ctx</th><td>{{.Status.NCtx}}</td></tr>
+<tr><th>n_gpu_layers</th><td>{{.Status.NGpuLayers}}</td></tr>
+</table></div>
+<p class=hint>the model name is what llama-server reports for itself, not what the config says it should be.</p>
+</section>
+<section class=card>
+<h2 class=card-title>allowed models <span class=badge>{{len .Status.Swappable}}</span></h2>
+{{if .Status.Swappable}}<div class=scroll><table><tr><th>file</th><th></th></tr>
+{{range .Status.Swappable}}<tr><td><code>{{.}}</code></td>
+<td><form method=post action=/models class=inline-form>
+<input type=hidden name=model_path value="{{.}}">
+<button class=btn>load this one</button></form></td></tr>{{end}}
+</table></div>
+{{else}}<div class=empty><div>no models allowlisted</div></div>{{end}}
+</section>
+{{end}}
+{{template "shellBottom"}}

cmd/mavweb/morning.html

@@ -8,7 +8,10 @@
 <div class=scroll><table class=mono>
 <tr><th>at<th>kind<th>what</tr>
 {{range .Items}}<tr>
-<td>{{.At.Format "15:04"}}</td>
+<!-- In his clock. A plan item's At is a calendar fact's Ts or a reminder's
+     FireTs, and the store hands both back as UTC, so the raw hour printed a
+     reminder here at an hour /reminders did not agree with (V-469). -->
+<td>{{.At.Local.Format "15:04"}}</td>
 <td class=gray>{{.Kind}}</td>
 <td>{{if .Uncertain}}<span class=hint title="relayed notification, not a calendar read">похоже,</span> {{end}}{{.Text}}</td>
 </tr>{{end}}

cmd/mavweb/morning_test.go

@@ -0,0 +1,49 @@
+package main
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+)
+
+// morningCore serves a canned checklist and day plan.
+type morningCore struct {
+	ipc.UnimplementedCoreAPI
+	status []ipc.MorningRoutineStatus
+	plan   ipc.DayPlan
+}
+
+func (c *morningCore) MorningStatus(context.Context) ([]ipc.MorningRoutineStatus, error) {
+	return c.status, nil
+}
+
+func (c *morningCore) DayPlan(context.Context) (ipc.DayPlan, error) { return c.plan, nil }
+
+func TestMorningRendersPlanTimesInLocalZone(t *testing.T) {
+	// A plan item's At is a calendar fact's Ts or a reminder's FireTs, and the
+	// store hands both back as UTC. Printed raw, /morning named an hour for a
+	// reminder that /reminders — which does call Local — disagreed with.
+	away := awayFromLocal()
+	at := time.Date(2026, 8, 1, 9, 0, 0, 0, time.UTC).In(away)
+	core := &morningCore{plan: ipc.DayPlan{
+		Date:  at,
+		Items: []ipc.DayPlanItem{{At: at, Text: "выпить таблетки", Kind: "reminder"}},
+	}}
+	w := httptest.NewRecorder()
+	handleMorning(w, httptest.NewRequest(http.MethodGet, "/morning", nil), core)
+	if w.Code != http.StatusOK {
+		t.Fatalf("status = %d, want 200", w.Code)
+	}
+	body := w.Body.String()
+	if !strings.Contains(body, at.Local().Format("15:04")) {
+		t.Errorf("plan item not rendered in his clock (%s): %s", at.Local().Format("15:04"), body)
+	}
+	if strings.Contains(body, at.In(away).Format("15:04")) {
+		t.Errorf("plan item rendered in the stored zone: %s", body)
+	}
+}

cmd/mavweb/notifications.go

@@ -0,0 +1,76 @@
+package main
+
+import (
+	_ "embed"
+	"log"
+	"net/http"
+	"strconv"
+
+	"github.com/kami/maven/internal/ipc"
+)
+
+//go:embed notifications.html
+var notificationsHTML string
+
+var notificationsTmpl = parsePage("notifications", notificationsHTML, nil)
+
+// deliveryRow is one outbox line, with every timestamp already formatted so
+// the template holds no date logic — same shape as taskRow.
+type deliveryRow struct {
+	Kind      string
+	Target    string
+	Channel   string
+	Status    string
+	Created   string
+	Completed string
+}
+
+func deliveryRows(as []ipc.DeliveryAttempt) []deliveryRow {
+	out := make([]deliveryRow, 0, len(as))
+	for _, a := range as {
+		target := a.Rule
+		if target == "" && a.ReminderID != 0 {
+			target = "reminder #" + strconv.FormatInt(a.ReminderID, 10)
+		}
+		row := deliveryRow{
+			Kind:    a.Kind,
+			Target:  target,
+			Channel: a.Channel,
+			Status:  a.Status,
+			Created: a.Created.Format("02.01 15:04"),
+		}
+		if a.Completed != nil {
+			row.Completed = a.Completed.Format("15:04")
+		}
+		out = append(out, row)
+	}
+	return out
+}
+
+func handleNotifications(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "notifications") {
+		return
+	}
+	ctx := r.Context()
+	nudges, err := core.RecentNudges(ctx, 50)
+	if err != nil {
+		log.Printf("notifications: %v", err)
+		http.Error(w, "notifications error: "+err.Error(), http.StatusBadGateway)
+		return
+	}
+	// The outbox, on the page that already answers "what did she send".
+	// A failed or dropped attempt is why she went quiet, and until now it was
+	// recorded and unreadable (Vikunja #390). Filter with ?status=dropped.
+	status := r.URL.Query().Get("status")
+	attempts, err := core.DeliveryAttempts(ctx, status, 50)
+	if err != nil {
+		// The nudge list is still worth showing, so this is a note on the page
+		// rather than a dead page.
+		log.Printf("notifications: delivery attempts: %v", err)
+	}
+	renderPage(w, notificationsTmpl, map[string]any{
+		"Nudges":   nudges,
+		"Attempts": deliveryRows(attempts),
+		"Status":   status,
+	})
+}

cmd/mavweb/pages.go

@@ -0,0 +1,205 @@
+package main
+
+import (
+	"cmp"
+	_ "embed"
+	"html/template"
+	"log"
+	"net/http"
+	"strings"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+)
+
+// The read-only pages: dash, history, trace, morning, events, and the voice
+// page mounted at "/". Each is GET-only, reads through CoreAPI and renders.
+// The write surfaces live next to their own handlers (tasks.go, tools.go,
+// routines.go, chat.go).
+
+//go:embed dash.html
+var dashHTML string
+
+//go:embed history.html
+var historyHTML string
+
+//go:embed trace.html
+var traceHTML string
+
+//go:embed morning.html
+var morningHTML string
+
+//go:embed events.html
+var eventsHTML string
+
+//go:embed voice.html
+var voiceHTML string
+
+//go:embed ecosystem.html
+var ecosystemHTML string
+
+// dashTmpl — the monitoring read surface, server-rendered from dash.html;
+// a small fetch loop refreshes the tables in place. html/template escapes the
+// user text in facts/nudges. Read-only: browses the append-only store via
+// CoreAPI, never writes — the store IS the audit trail, this just shows it.
+var dashTmpl = parsePage("dash", dashHTML, nil)
+
+var historyTmpl = parsePage("history", historyHTML, nil)
+
+var traceTmpl = parsePage("trace", traceHTML, template.FuncMap{
+	"fmtTime": func(t *time.Time) string {
+		if t == nil || t.IsZero() {
+			return "—"
+		}
+		return t.Format("15:04:05")
+	},
+	"join": strings.Join,
+})
+
+// morningTmpl — read-only view of today's checklist state per configured
+// morning routine (internal/morning). Same shape as trace.html: a plain
+// server-rendered page, refreshed on reload — no live-update loop, since
+// checklist state changes on the scale of minutes, not seconds.
+var morningTmpl = parsePage("morning", morningHTML, nil)
+
+// eventsTmpl — the unified intake journal (Vikunja #283), read-only. Same
+// shape as trace.html and morning.html: server-rendered, refreshed on reload.
+var eventsTmpl = parsePage("events", eventsHTML, nil)
+
+var voiceTmpl = parsePage("voice", voiceHTML, nil)
+
+// ecosystemTmpl — read-only view of the Nexus/Praxis/Hexis siblings, whose only
+// human surface is here (they ship no web UI of their own).
+var ecosystemTmpl = parsePage("ecosystem", ecosystemHTML, nil)
+
+func handleDash(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "dash") {
+		return
+	}
+	ctx := r.Context()
+	pres, err1 := core.Presence(ctx)
+	facts, err2 := core.RecentFacts(ctx, 50)
+	nudges, err3 := core.RecentNudges(ctx, 50)
+	notes, err4 := core.RecentNotes(ctx, 50)
+	if err := cmp.Or(err1, err2, err3, err4); err != nil {
+		log.Printf("dash: %v", err)
+		http.Error(w, "core read failed", http.StatusBadGateway)
+		return
+	}
+	renderPage(w, dashTmpl, struct {
+		Presence ipc.Presence
+		Facts    []ipc.Fact
+		Nudges   []ipc.Nudge
+		Notes    []ipc.Note
+	}{pres, facts, nudges, notes})
+}
+
+func handleHistory(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "history") {
+		return
+	}
+	facts, err := core.RecentFacts(r.Context(), 200)
+	if err != nil {
+		log.Printf("history: %v", err)
+		http.Error(w, "core read failed", http.StatusBadGateway)
+		return
+	}
+	renderPage(w, historyTmpl, struct {
+		Facts []ipc.Fact
+	}{facts})
+}
+
+func handleTrace(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "trace") {
+		return
+	}
+	trace, err := core.TickTrace(r.Context())
+	if err != nil {
+		log.Printf("trace: %v", err)
+		http.Error(w, "core read failed", http.StatusBadGateway)
+		return
+	}
+	// The turn records share this page rather than getting one of their own
+	// (V-564). Both answer the same question, who won and who lost and why, and
+	// one is about nudges while the other is about utterances. A read failure
+	// here is not fatal to the page. The rule trace above it still renders, and
+	// a daemon too old to know the method is the ordinary case during a rolling
+	// deploy.
+	turns, err := core.TurnDecisions(r.Context(), 25)
+	if err != nil {
+		log.Printf("trace: turn decisions: %v", err)
+	}
+	renderPage(w, traceTmpl, traceData{Tick: trace, Turns: turns})
+}
+
+// traceData — what trace.html renders: the last tick's rule arbitration and the
+// last turns' claim arbitration.
+type traceData struct {
+	Tick  ipc.TickTrace
+	Turns []ipc.TurnDecision
+}
+
+// morningView — what /morning renders: today's plan on top, the checklist
+// state under it. PlanErr is set instead of Plan when the core could not build
+// a plan, so the page says so rather than showing an empty day.
+type morningView struct {
+	Plan     *ipc.DayPlan
+	PlanErr  string
+	Routines []ipc.MorningRoutineStatus
+}
+
+func handleMorning(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "morning") {
+		return
+	}
+	ctx := r.Context()
+	status, err := core.MorningStatus(ctx)
+	if err != nil {
+		log.Printf("morning: %v", err)
+		http.Error(w, "core read failed", http.StatusBadGateway)
+		return
+	}
+	view := morningView{Routines: status}
+	// The day plan (#128) shows on this page because it is the same question at
+	// a different scale. A plan read that fails must not take the checklist
+	// down with it — the page degrades to what it had before.
+	plan, err := core.DayPlan(ctx)
+	if err != nil {
+		log.Printf("morning: day plan: %v", err)
+		view.PlanErr = err.Error()
+	} else {
+		view.Plan = &plan
+	}
+	renderPage(w, morningTmpl, view)
+}
+
+// eventsView — what /events renders. Err is set instead of Events when the
+// core could not serve the journal, so the page says why rather than showing an
+// empty intake and implying nothing arrived.
+type eventsView struct {
+	Events []ipc.IntakeEvent
+	Err    string
+}
+
+// eventsPageLimit — how many envelopes the page shows. The ring holds more; a
+// page is for scanning what just happened, not for archaeology.
+const eventsPageLimit = 200
+
+func handleEvents(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "intake journal") {
+		return
+	}
+	var view eventsView
+	evs, err := core.RecentEvents(r.Context(), eventsPageLimit)
+	if err != nil {
+		log.Printf("events: %v", err)
+		view.Err = err.Error()
+	} else {
+		view.Events = evs
+	}
+	renderPage(w, eventsTmpl, view)
+}
+
+func handleVoice(w http.ResponseWriter, r *http.Request) {
+	renderPage(w, voiceTmpl, nil)
+}

cmd/mavweb/passkey.html

@@ -0,0 +1,58 @@
+{{template "shellTop" "passkey"}}
+<h1>Passkey</h1>
+<p class=hint>Enroll a passkey once, then assert it to unlock destructive actions (tool enable) for a few minutes.</p>
+<div class=flex gap-2>
+<button class=btn onclick=enroll()>enroll passkey</button>
+<button class=btn onclick=assert()>assert (step-up)</button>
+<button class=btn onclick=rewrapKey()>rewrite cold-start key</button>
+<a href=/tools><button class=btn-primary>→ tools</button></a>
+</div>
+<p class=hint>Rewriting the cold-start key points it at the passkey you assert next. Every other enrolled passkey stops being able to unlock a cold-booted daemon.</p>
+<div id=msg></div>
+{{template "shellBottom"}}
+<script>
+const b64u=b=>btoa(String.fromCharCode(...new Uint8Array(b))).replace(/\+/g,'-').replace(/\//g,'_').replace(/=+$/,'');
+const ub64=s=>{s=s.replace(/-/g,'+').replace(/_/g,'/');const b=atob(s),a=new Uint8Array(b.length);for(let i=0;i<b.length;i++)a[i]=b.charCodeAt(i);return a;};
+const say=(t,ok)=>{const m=document.getElementById('msg');m.textContent=t;m.className=ok?'msg msg-ok':'msg msg-err';};
+async function enroll(){try{
+ const {challenge,options}=await (await fetch('/auth/webauthn/register/begin')).json();
+ options.challenge=ub64(options.challenge);
+ options.user.id=ub64(options.user.id);
+ const c=await navigator.credentials.create({publicKey:options});
+ const r=await fetch('/auth/webauthn/register/finish',{method:'POST',headers:{'content-type':'application/json'},
+  body:JSON.stringify({challenge,credential:{id:c.id,type:c.type,response:{
+   clientDataJSON:b64u(c.response.clientDataJSON),attestationObject:b64u(c.response.attestationObject)}}})});
+ if(!r.ok){say('enroll failed: '+await r.text(),false);return;}
+ // The wrapped key can only be written from an assertion: PRF results are
+ // not produced at create() time on most authenticators. Enrolment reports
+ // whether PRF is available at all so he is not told cold-start works when
+ // it cannot.
+ const ext=c.getClientExtensionResults?c.getClientExtensionResults():{};
+ const prfOK=!!(ext.prf&&ext.prf.enabled);
+ say(prfOK?'enrolled ✓ — now assert once to write the cold-start key':
+  'enrolled ✓ — but this authenticator has no PRF: cold-start unlock unavailable',true);
+}catch(e){say('enroll error: '+e,false);}}
+async function assert(explicit){try{
+ const {challenge,options}=await (await fetch('/auth/webauthn/assert/begin')).json();
+ options.challenge=ub64(options.challenge);
+ const c=await navigator.credentials.get({publicKey:options});
+ // The PRF result is the cold-start secret. It never touches localStorage
+ // and is posted once, over the same request as the assertion.
+ const ext=c.getClientExtensionResults?c.getClientExtensionResults():{};
+ const prf=ext.prf&&ext.prf.results&&ext.prf.results.first?b64u(ext.prf.results.first):'';
+ const r=await fetch('/auth/webauthn/assert/finish',{method:'POST',headers:{'content-type':'application/json'},
+  body:JSON.stringify({challenge,prf,explicit:!!explicit,credential:{id:c.id,type:c.type,response:{
+   clientDataJSON:b64u(c.response.clientDataJSON),authenticatorData:b64u(c.response.authenticatorData),
+   signature:b64u(c.response.signature)}}})});
+ if(!r.ok){say('assert failed: '+await r.text(),false);return;}
+ if(!prf){say('stepped up ✓ — no PRF from this authenticator, so cold-start unlock stayed unavailable',true);return;}
+ say(explicit?'stepped up ✓ — cold-start key now points at this passkey':
+  'stepped up ✓ — enable tools now',true);
+}catch(e){say('assert error: '+e,false);}}
+// Rewriting the wrapped key is a separate gesture, never a side effect of a
+// step-up. Only this button sets explicit, and only explicit lets the daemon
+// replace a blob that already exists.
+async function rewrapKey(){
+ if(!confirm('Rewrite the cold-start key under the passkey you are about to assert? Every other enrolled passkey stops being able to unlock a cold-booted daemon.'))return;
+ await assert(true);}
+</script>

cmd/mavweb/reminders.go

@@ -0,0 +1,74 @@
+package main
+
+import (
+	_ "embed"
+	"encoding/json"
+	"log"
+	"net/http"
+	"strings"
+
+	"github.com/kami/maven/internal/ipc"
+)
+
+//go:embed reminders.html
+var remindersHTML string
+
+var remindersTmpl = parsePage("reminders", remindersHTML, nil)
+
+// reminderRow is one line on /reminders, with the payload unwrapped and both
+// timestamps already in his clock.
+//
+// The page rendered `{{.Payload}}` and the UTC instant, so a reminder read
+// `{"text":"выпить таблетки"}` and fired an hour off what he was told
+// (Vikunja #469). Neither is a formatting nicety: the envelope is an internal
+// shape he never chose, and a time on a page he reads is the time on his wall.
+type reminderRow struct {
+	Created string
+	Fires   string
+	Status  string
+	Text    string
+}
+
+// reminderText unwraps the {"text":...} payload the router writes.
+//
+// A copy of store.ReminderText rather than a call to it, because mavweb is one
+// of the pure-Go daemons and internal/store carries the CGO sqlite driver. The
+// ipc DTO is decoupled from the store on purpose, so the unwrap belongs to
+// whoever renders it. Payload that is not that shape is shown as he said it.
+func reminderText(payload string) string {
+	var m map[string]any
+	if err := json.Unmarshal([]byte(payload), &m); err == nil {
+		if t, ok := m["text"]; ok {
+			if s, isStr := t.(string); isStr && s != "" {
+				return s
+			}
+		}
+	}
+	return strings.TrimSpace(payload)
+}
+
+func reminderRows(rs []ipc.Reminder) []reminderRow {
+	out := make([]reminderRow, 0, len(rs))
+	for _, r := range rs {
+		out = append(out, reminderRow{
+			Created: r.CreatedTs.Local().Format("02 Jan 15:04"),
+			Fires:   r.FireTs.Local().Format("02 Jan 15:04"),
+			Status:  r.Status,
+			Text:    reminderText(r.Payload),
+		})
+	}
+	return out
+}
+
+func handleReminders(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "reminders") {
+		return
+	}
+	reminders, err := core.ListReminders(r.Context(), 50)
+	if err != nil {
+		log.Printf("reminders: %v", err)
+		http.Error(w, "reminders error: "+err.Error(), http.StatusBadGateway)
+		return
+	}
+	renderPage(w, remindersTmpl, map[string]any{"Reminders": reminderRows(reminders)})
+}

cmd/mavweb/routines.go

@@ -0,0 +1,200 @@
+package main
+
+import (
+	"context"
+	_ "embed"
+	"errors"
+	"fmt"
+	"log"
+	"net/http"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/pattern"
+	"github.com/kami/maven/internal/webauthn"
+)
+
+//go:embed routines.html
+var routinesHTML string
+
+// routinesTmpl — the proposed-routine review surface. One row per thing maven
+// noticed, in her words, with at most two actions: accept or dismiss.
+var routinesTmpl = parsePage("routines", routinesHTML, nil)
+
+// routineView is one line on the page: what maven noticed, in her words, and
+// how long ago she noticed it. A view model, not a database row — the template
+// never formats an interval or a timestamp itself.
+type routineView struct {
+	ID      int64
+	Phrase  string
+	Noticed string
+}
+
+// handleRoutines serves the routine review surface (GET) and answers a
+// proposal (POST id + action=accept|dismiss).
+//
+// Accept is gated at step-up, the same tier as enabling a tool: saying yes
+// hands the trigger loop a new standing reason to speak to the human, so it
+// moves the boundary and only an authed surface may do it. Dismiss is not
+// gated — it only ever removes a reason to speak, so the worst a weaker caller
+// can do is make maven quieter.
+func handleRoutines(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, session *webauthn.PasskeySession, requireStepUp bool) {
+	if !requireCore(w, core, "routines") {
+		return
+	}
+	ctx := r.Context()
+	var msg string
+	if r.Method == http.MethodPost {
+		var ok bool
+		if msg, ok = applyRoutinePost(w, r, core, session, requireStepUp); !ok {
+			return
+		}
+	}
+	proposed, err := core.ListProposedRoutines(ctx)
+	if err != nil {
+		log.Printf("routines: list: %v", err)
+		http.Error(w, "routines error: "+err.Error(), http.StatusBadGateway)
+		return
+	}
+	renderPage(w, routinesTmpl, struct {
+		Msg      string
+		Proposed []routineView
+	}{msg, toRoutineViews(proposed)})
+}
+
+// applyRoutinePost performs one write and returns the message to show. Unlike
+// the task form, a bad request here is an HTTP status rather than an inline
+// note, so the second return says whether the response was already written.
+func applyRoutinePost(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, session *webauthn.PasskeySession, requireStepUp bool) (string, bool) {
+	ctx := r.Context()
+	action := r.FormValue("action")
+	// "seed" is the one action with no routine to act on — it is what
+	// MAKES a routine (Vikunja #518), so it runs before the id parse. It
+	// lives on this route rather than a page of its own because it is
+	// already the step-up-gated surface for this table, and a second gated
+	// surface is a second thing to get wrong.
+	if action == "seed" {
+		if !stepUpGate(w, session, requireStepUp) {
+			return "", false
+		}
+		out, err := seedRoutineEvent(ctx, core, r)
+		if err != nil {
+			log.Printf("routines: seed: %v", err)
+			http.Error(w, "seed failed: "+err.Error(), http.StatusBadGateway)
+			return "", false
+		}
+		return out, true
+	}
+
+	idStr := r.FormValue("id")
+	var rid int64
+	if n, _ := fmt.Sscanf(idStr, "%d", &rid); n != 1 {
+		http.Error(w, "invalid id", http.StatusBadRequest)
+		return "", false
+	}
+	switch action {
+	case "accept":
+		if !stepUpGate(w, session, requireStepUp) {
+			return "", false
+		}
+		if err := acceptRoutine(ctx, core, rid); err != nil {
+			log.Printf("routines: accept %d: %v", rid, err)
+			http.Error(w, "accept failed: "+err.Error(), http.StatusBadGateway)
+			return "", false
+		}
+		return "accepted routine — maven will remind you", true
+	case "dismiss":
+		if err := core.DismissProposedRoutine(ctx, rid); err != nil {
+			log.Printf("routines: dismiss %d: %v", rid, err)
+			http.Error(w, "dismiss failed: "+err.Error(), http.StatusBadGateway)
+			return "", false
+		}
+		return "dismissed routine", true
+	default:
+		http.Error(w, "unknown action", http.StatusBadRequest)
+		return "", false
+	}
+}
+
+// toRoutineViews turns the wire rows into view models. The phrase comes from
+// pattern.PhraseRoutine so the page says the same thing maven's voice says.
+func toRoutineViews(rs []ipc.ProposedRoutine) []routineView {
+	out := make([]routineView, 0, len(rs))
+	for _, r := range rs {
+		p := pattern.ProposedRoutine{Action: r.Action, Object: r.Object, IntervalDays: r.IntervalDays}
+		noticed := "just now"
+		if r.CreatedTs > 0 {
+			noticed = time.Since(time.UnixMilli(r.CreatedTs)).Round(time.Minute).String() + " ago"
+		}
+		out = append(out, routineView{ID: r.ID, Phrase: pattern.PhraseRoutine(&p), Noticed: noticed})
+	}
+	return out
+}
+
+// acceptRoutine marks a proposal accepted. This page is the ONLY surface that
+// may do it (Vikunja #367): accepting gives the tick loop a standing new
+// reason to speak, which DESIGN.md puts at layer 3, and the button here is
+// behind step-up. Voice can park the question and dismiss, never accept.
+func acceptRoutine(ctx context.Context, core ipc.CoreAPI, id int64) error {
+	proposed, err := core.ListProposedRoutines(ctx)
+	if err != nil {
+		return err
+	}
+	var found *ipc.ProposedRoutine
+	for i := range proposed {
+		if proposed[i].ID == id {
+			found = &proposed[i]
+			break
+		}
+	}
+	if found == nil {
+		return errors.New("no such proposed routine")
+	}
+
+	// No reminder is created here. Accepting only flips the status; the tick
+	// loop reads accepted routines and nudges on the interval (Vikunja #366).
+	// The old code made a one-shot reminder, so a non-weekly routine fired
+	// once and then went quiet forever.
+	return core.AcceptProposedRoutine(ctx, id)
+}
+
+// seedRoutineEvent drives one backdated fact write through core (Vikunja #518),
+// so the pattern detector can be exercised against a running daemon instead of
+// over real days. Refused unless mavend was started with -allow-seed; on an
+// ordinary box the error says so and nothing is written.
+//
+// Takes "ago" rather than an absolute timestamp — hours before now, as a float
+// so a QA sitting can space four seeds three hours apart without doing clock
+// arithmetic. The detector's floor is two hours, and "0" is a legal answer
+// meaning now.
+func seedRoutineEvent(ctx context.Context, core ipc.CoreAPI, r *http.Request) (string, error) {
+	key := strings.TrimSpace(r.FormValue("key"))
+	value := strings.TrimSpace(r.FormValue("value"))
+	if key == "" || value == "" {
+		return "", errors.New("seed needs a key and a value")
+	}
+	agoHours, err := strconv.ParseFloat(strings.TrimSpace(r.FormValue("ago")), 64)
+	if err != nil {
+		return "", fmt.Errorf("seed: bad ago (hours before now): %w", err)
+	}
+	if agoHours < 0 {
+		return "", errors.New("seed: ago is hours BEFORE now, so it cannot be negative")
+	}
+	resp, err := core.SeedEvent(ctx, ipc.SeedEventReq{
+		Key:   key,
+		Value: value,
+		Ts:    time.Now().Add(-time.Duration(agoHours * float64(time.Hour))),
+	})
+	if err != nil {
+		return "", err
+	}
+	if !resp.Extracted {
+		return fmt.Sprintf("wrote fact %d, but %q is not in the action lexicon — no event, no pattern", resp.FactID, value), nil
+	}
+	if !resp.Proposed {
+		return fmt.Sprintf("seeded %s/%s (fact %d, event %d) — not enough yet to propose", resp.Action, resp.Object, resp.FactID, resp.EventID), nil
+	}
+	return fmt.Sprintf("seeded %s/%s and PROPOSED routine %d, every %.1f days", resp.Action, resp.Object, resp.RoutineID, resp.IntervalDays), nil
+}

cmd/mavweb/shell.go

@@ -0,0 +1,195 @@
+package main
+
+import (
+	_ "embed"
+	"html/template"
+	"log"
+	"net/http"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/webauthn"
+)
+
+// shellHTML — the shell partial every page is wrapped in: "shellTop", the
+// "sidebar" it calls, and "shellBottom". It used to be two Go string constants
+// with the sidebar assembled by a strings.Builder, which is the one piece of
+// markup that was still concatenated in Go.
+//
+// Two template pieces wrap every page:
+//
+//	{{template "shellTop" "<page-key>"}}  ← opens <html>, topbar, sidebar, content
+//	{{template "shellBottom"}}            ← closes content, inspector, </html>
+//
+// The page-key argument highlights the active sidebar link and sets breadcrumbs.
+//
+//go:embed shell.html
+var shellHTML string
+
+// sidebarSections maps sidebar section → page entries {label, url, icon}
+var sidebarSections = []struct {
+	Label string
+	Pages []struct{ Label, URL, Key string }
+}{
+	{
+		Label: "Workspace",
+		Pages: []struct{ Label, URL, Key string }{
+			{Label: "Dashboard", URL: "/dash", Key: "dash"},
+		},
+	},
+	{
+		Label: "Infrastructure",
+		Pages: []struct{ Label, URL, Key string }{
+			{Label: "History", URL: "/history", Key: "history"},
+		},
+	},
+	{
+		Label: "Automation",
+		Pages: []struct{ Label, URL, Key string }{
+			{Label: "Rule Trace", URL: "/trace", Key: "trace"},
+			{Label: "Notifications", URL: "/notifications", Key: "notifications"},
+			{Label: "Tasks", URL: "/tasks", Key: "tasks"},
+			{Label: "Reminders", URL: "/reminders", Key: "reminders"},
+			{Label: "Routines", URL: "/routines", Key: "routines"},
+			{Label: "Morning", URL: "/morning", Key: "morning"},
+			{Label: "Intake", URL: "/events", Key: "events"},
+		},
+	},
+	{
+		Label: "Ecosystem",
+		Pages: []struct{ Label, URL, Key string }{
+			{Label: "Siblings", URL: "/ecosystem", Key: "ecosystem"},
+		},
+	},
+	{
+		Label: "AI",
+		Pages: []struct{ Label, URL, Key string }{
+			{Label: "Chat", URL: "/chat", Key: "chat"},
+			{Label: "Voice", URL: "/", Key: "voice"},
+		},
+	},
+	{
+		Label: "Settings",
+		Pages: []struct{ Label, URL, Key string }{
+			{Label: "Tools", URL: "/tools", Key: "tools"},
+			{Label: "Model", URL: "/models", Key: "models"},
+			{Label: "Passkey", URL: "/auth/passkey", Key: "passkey"},
+		},
+	},
+}
+
+// pageChrome is the per-page title and ethos-icons.svg symbol id, keyed by the
+// page key a page hands to shellTop. One table rather than two parallel
+// switches, so a new page cannot end up with a title and no icon.
+var pageChrome = map[string]struct{ Title, Icon string }{
+	"dash":          {"Dashboard", "i-grid"},
+	"history":       {"History", "i-clock"},
+	"trace":         {"Rule Trace", "i-wave"},
+	"notifications": {"Notifications", "i-bell"},
+	"tasks":         {"Tasks", "i-grid"},
+	"reminders":     {"Reminders", "i-calendar"},
+	"routines":      {"Routines", "i-repeat"},
+	"morning":       {"Morning Routines", "i-calendar"},
+	"events":        {"Intake", "i-download"},
+	"chat":          {"Chat", "i-message"},
+	"voice":         {"Voice", "i-mic"},
+	"ecosystem":     {"Ecosystem", "i-grid"},
+	"tools":         {"Tools", "i-settings"},
+	"models":        {"Resident Model", "i-wave"},
+	"passkey":       {"Passkey", "i-lock"},
+}
+
+// pageIcon returns the ethos-icons.svg symbol id for the given page. The
+// sidebar template wraps it in the <use> reference.
+func pageIcon(key string) string {
+	if c, ok := pageChrome[key]; ok {
+		return c.Icon
+	}
+	return "i-search"
+}
+
+// pageTitle returns the human-readable page title for the given key. An
+// unknown key renders as itself rather than as a blank crumb.
+func pageTitle(key string) string {
+	if c, ok := pageChrome[key]; ok {
+		return c.Title
+	}
+	return key
+}
+
+// shellFuncs returns the FuncMap shared by every server-rendered page template.
+func shellFuncs() template.FuncMap {
+	return template.FuncMap{
+		"pageTitle":       pageTitle,
+		"pageIcon":        pageIcon,
+		"sidebarSections": func() any { return sidebarSections },
+		"ago": func(t time.Time) string {
+			if t.IsZero() {
+				return "never"
+			}
+			return time.Since(t).Round(time.Second).String() + " ago"
+		},
+	}
+}
+
+// parsePage parses one server-rendered page: the shell partial plus the page's
+// own embedded markup, under the shared FuncMap. extra adds page-local
+// functions (/tools needs capability lookups, /trace a time format) and may be
+// nil.
+//
+// The name is also the page's log label, so a render failure says which page.
+func parsePage(name, body string, extra template.FuncMap) *template.Template {
+	funcs := shellFuncs()
+	for k, v := range extra {
+		funcs[k] = v
+	}
+	return template.Must(template.New(name).Funcs(funcs).Parse(shellHTML + body))
+}
+
+// renderPage writes one page. Every handler sent the same content type and
+// logged the same way on failure; the header is already written by then, so a
+// render error can only be logged, never reported.
+func renderPage(w http.ResponseWriter, t *template.Template, data any) {
+	w.Header().Set("Content-Type", "text/html; charset=utf-8")
+	if err := t.Execute(w, data); err != nil {
+		log.Printf("%s render: %v", t.Name(), err)
+	}
+}
+
+// requireCore answers whether the surface has a core to read. mavweb runs
+// without -core (voice-only), and every page that needs mavend says so with a
+// 503 naming itself rather than a blank error.
+func requireCore(w http.ResponseWriter, core ipc.CoreAPI, surface string) bool {
+	if core == nil {
+		http.Error(w, surface+" disabled (no -core)", http.StatusServiceUnavailable)
+		return false
+	}
+	return true
+}
+
+// stepUpGate reports whether the caller may proceed through the AuthStepUp
+// gate, writing the 403 itself when it may not. See stepUpOK for the policy.
+func stepUpGate(w http.ResponseWriter, session *webauthn.PasskeySession, requireStepUp bool) bool {
+	if stepUpOK(session, requireStepUp) {
+		return true
+	}
+	http.Error(w, "step-up required: assert a passkey first", http.StatusForbidden)
+	return false
+}
+
+// stepUpOK is the single decision point for the AuthStepUp gate shared by
+// POST /tools and POST /api/revert.
+//
+// A nil session means WebAuthn is not configured (-webauthn-origin /
+// -webauthn-rpid unset), so step-up can never be asserted — not merely unmet.
+// The default is therefore fail-OPEN: gating on an unassertable session would
+// 403 those surfaces permanently. In that mode the actions rest on the
+// transport-level auth in front of mavweb (wg+nginx+auth), and main logs a
+// startup warning naming them. With -require-stepup the same situation fails
+// CLOSED instead: no assertable step-up ⇒ deny.
+func stepUpOK(session *webauthn.PasskeySession, requireStepUp bool) bool {
+	if session == nil {
+		return !requireStepUp
+	}
+	return session.IsStepUp()
+}

cmd/mavweb/static/ui.css

@@ -702,6 +702,11 @@ details[open] > summary { margin-bottom: var(--space-1); }
 .chat-form { display: flex; gap: var(--space-2); }
 .chat-form input { flex: 1; }
 .chat-scroll { max-height: 60vh; overflow-y: auto; margin-bottom: var(--space-4); }
+/* The correction gesture (V-630). Wraps on a phone rather than scrolling: it is
+   one row of small buttons, and a gesture that has to be panned to is not one. */
+.chat-correct { display: flex; flex-wrap: wrap; align-items: center; gap: var(--space-1);
+  padding: 0 var(--space-3) var(--space-2); margin-top: calc(-1 * var(--space-1)); margin-bottom: var(--space-2); }
+.chat-correct-label { font-size: var(--fs-xs); color: var(--text-machine); margin-left: var(--space-2); }
 
 /* ── Key-value grid ── */
 .kv { display: grid; grid-template-columns: auto 1fr; gap: var(--space-1) var(--space-3); font-size: var(--fs-sm); }

cmd/mavweb/tasks.go

@@ -0,0 +1,412 @@
+package main
+
+import (
+	"context"
+	_ "embed"
+	"errors"
+	"fmt"
+	"log"
+	"net/http"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/tasks"
+)
+
+//go:embed tasks.html
+var tasksHTML string
+
+var tasksTmpl = parsePage("tasks", tasksHTML, nil)
+
+// now — the wall clock, indirected so the task page can be rendered at a fixed
+// instant in a test. internal/tasks is pure and the daemon path already ranks
+// through a clock it is handed; the page had no reason to be the one surface
+// that could only be tested at whatever time it happened to run.
+var now = time.Now
+
+// resolvedShown — how many finished tasks the page renders. The list is
+// history, it only grows, and the rows below the first screen are read by
+// nobody.
+const resolvedShown = 50
+
+// taskRow is one line on /tasks, with every timestamp already formatted so the
+// template holds no date logic.
+type taskRow struct {
+	ID         int64
+	Text       string
+	Source     string
+	Evidence   string
+	Status     string
+	Due        string
+	Created    string
+	Resolved   string
+	ResolvedBy string
+	// DueValue and Weight are the raw values the edit form posts back
+	// (Vikunja #509). Due above is for reading and says "—" for no date; a
+	// date input needs "2026-08-07" or the empty string.
+	DueValue string
+	Weight   int
+	// Why — the ranker's reason for this row's position (Vikunja #129), in
+	// Russian, empty when nothing distinguished the task. Blank is the honest
+	// rendering: he never said this one mattered more.
+	Why string
+}
+
+// rowOf renders one wire task into the shared read-only columns. The two call
+// sites below add what only they need: the live rows carry the edit form's raw
+// values and the ranker's reason, the resolved rows carry neither.
+func rowOf(t ipc.Task) taskRow {
+	return taskRow{
+		ID: t.ID, Text: t.Text, Source: t.Source, Evidence: t.Evidence,
+		Status: t.Status, Created: fmtTaskTime(&t.CreatedTs),
+		Due: fmtTaskDate(t.Due), Resolved: fmtTaskTime(t.Resolved),
+		ResolvedBy: t.ResolvedBy,
+	}
+}
+
+// handleTasks serves the task review surface (GET) and the five writes it
+// offers (POST): add, edit, confirm, done, drop.
+//
+// Not step-up gated, unlike /tools and /routines, and the difference is the
+// point: enabling a tool defines argv Maven will execute, and accepting a
+// routine hands the tick loop a new standing reason to interrupt him. A task is
+// neither — nothing in the tick loop reads the tasks table, so the worst a
+// weaker caller can do here is write a line onto a list he reads himself. It
+// still sits behind whatever transport auth fronts mavweb, like every other
+// page.
+//
+// "edit" was re-argued on the same terms rather than inheriting the exemption
+// (Vikunja #509), and it stays ungated. It rewrites a line on a list he reads
+// himself, the same blast radius "drop" already has on this page, and the store
+// refuses the two edits that would cost something: a resolved task keeps the
+// text it was finished under, and a text collision with another live row is
+// named instead of merged.
+//
+// "confirm" is the only interesting move: it promotes a candidate Maven derived
+// from something she read into work he owns. That review step is why derived
+// tasks are captured as candidates in the first place.
+func handleTasks(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI) {
+	if !requireCore(w, core, "tasks") {
+		return
+	}
+	ctx := r.Context()
+	var msg, errMsg string
+	if r.Method == http.MethodPost {
+		var err error
+		msg, err = applyTaskPost(ctx, core, r)
+		if err != nil {
+			log.Printf("tasks: %v", err)
+			errMsg = err.Error()
+		}
+	}
+
+	all, err := core.ListTasks(ctx, "")
+	if err != nil {
+		log.Printf("tasks: list: %v", err)
+		http.Error(w, "tasks error: "+err.Error(), http.StatusBadGateway)
+		return
+	}
+	// Live rows are ordered by the same ranker the spoken list uses, so the page
+	// and the voice reply can never disagree about what comes first. Resolved
+	// rows keep store order (newest first) — ranking finished work is pointless.
+	var live []tasks.Item
+	var resolved []taskRow
+	resolvedTotal := 0
+	for _, t := range all {
+		switch t.Status {
+		case "candidate", "open":
+			live = append(live, tasks.Item{
+				ID: t.ID, Text: t.Text, Status: t.Status,
+				Created: t.CreatedTs, Due: t.Due, Weight: t.Weight,
+			})
+		default:
+			resolvedTotal++
+			// Finished work is history, and the history only grows. The page
+			// showed every row that ever existed, which is a page that gets
+			// slower every month for a section nobody reads past the top of.
+			if len(resolved) >= resolvedShown {
+				continue
+			}
+			resolved = append(resolved, rowOf(t))
+		}
+	}
+	byID := make(map[int64]ipc.Task, len(all))
+	for _, t := range all {
+		byID[t.ID] = t
+	}
+	var cands, open []taskRow
+	for _, r := range tasks.Rank(live, now()) {
+		t := byID[r.ID]
+		row := rowOf(t)
+		row.DueValue = fmtTaskDateValue(t.Due)
+		row.Weight = t.Weight
+		row.Why = r.Reason
+		if t.Status == "candidate" {
+			// A candidate's due date is Maven's reading of a mail, so its
+			// ranking reason is not shown as if he had set a priority.
+			row.Why = ""
+			cands = append(cands, row)
+		} else {
+			open = append(open, row)
+		}
+	}
+	renderPage(w, tasksTmpl, struct {
+		Msg, Err     string
+		Stalls       []tasks.Stall
+		Candidates   []taskRow
+		Open         []taskRow
+		Resolved     []taskRow
+		ResolvedMore bool
+	}{msg, errMsg, tasks.Stalls(live, now()), cands, open, resolved, resolvedTotal > len(resolved)})
+}
+
+// applyTaskPost performs one write and returns the message to show. A bad
+// request returns an error, which the page renders inline rather than as a
+// bare 400 — this is a form surface, not an API.
+func applyTaskPost(ctx context.Context, core ipc.CoreAPI, r *http.Request) (string, error) {
+	action := r.FormValue("action")
+	if action == "add" {
+		text := strings.TrimSpace(r.FormValue("text"))
+		if text == "" {
+			return "", errors.New("empty task text")
+		}
+		req := ipc.CaptureTaskReq{Text: text, Source: "tap:web", Status: "open", Ts: now()}
+		wgt, err := formWeight(r)
+		if err != nil {
+			return "", err
+		}
+		req.Weight = wgt
+		due, err := formDue(r, now())
+		if err != nil {
+			return "", err
+		}
+		req.Due = due
+		resp, err := core.CaptureTask(ctx, req)
+		if err != nil {
+			return "", err
+		}
+		if resp.Promoted {
+			return "confirmed a candidate maven had found", nil
+		}
+		if !resp.Created {
+			return "already on the list", nil
+		}
+		return "added task", nil
+	}
+
+	id, err := strconv.ParseInt(r.FormValue("id"), 10, 64)
+	if err != nil {
+		return "", errors.New("invalid id")
+	}
+
+	if action == "promote" {
+		msg, err := promoteCandidate(ctx, core, r, id)
+		if err != nil {
+			return "", err
+		}
+		return msg, nil
+	}
+
+	if action == "edit" {
+		// The three fields capture set, and only those (Vikunja #509). Status
+		// is not editable here: that ladder is one-way and has its own buttons.
+		text := strings.TrimSpace(r.FormValue("text"))
+		if text == "" {
+			return "", errors.New("empty task text")
+		}
+		wgt, err := formWeight(r)
+		if err != nil {
+			return "", err
+		}
+		due, err := formDue(r, now())
+		if err != nil {
+			return "", err
+		}
+		switch err := core.EditTask(ctx, id, text, due, wgt); {
+		case err == nil:
+			return "saved task", nil
+		case errors.Is(err, ipc.ErrTaskDuplicate):
+			// Naming the collision instead of merging: two live rows carry two
+			// provenances, and picking one is not the page's call.
+			return "", errors.New("another open task already says this — drop one of the two")
+		case errors.Is(err, ipc.ErrTaskResolved):
+			return "", errors.New("a resolved task keeps the text it was finished under")
+		default:
+			return "", err
+		}
+	}
+
+	var status, msg string
+	switch action {
+	case "confirm":
+		status, msg = "open", "confirmed task"
+	case "done":
+		status, msg = "done", "task done"
+	case "drop":
+		status, msg = "dropped", "dropped task"
+	default:
+		return "", fmt.Errorf("unknown action %q", action)
+	}
+	if err := core.SetTaskStatus(ctx, id, status, now(), "tap:web"); err != nil {
+		return "", statusWriteErr(err)
+	}
+	return msg, nil
+}
+
+// errNoDoneWhen — the refusal has to name what is missing, or the button looks
+// broken. The field it asks for arrives with the intake form (Vikunja #511).
+var errNoDoneWhen = errors.New("write a definition of done before confirming this candidate")
+
+// statusWriteErr translates a SetTaskStatus failure into what the page says.
+func statusWriteErr(err error) error {
+	if errors.Is(err, ipc.ErrTaskNoDoneWhen) {
+		return errNoDoneWhen
+	}
+	return err
+}
+
+// promoteCandidate turns a candidate into open work with the three things the
+// board needs (Vikunja #511): a definition of done, an optional blocker, and an
+// optional date.
+//
+// The definition of done is required, and the refusal is the store's — this
+// only reaches it in a readable order. The blocker is a NAME here and an entity
+// id in the row: identity lives in Nexus, so the name is resolved first and a
+// name Nexus cannot resolve stops the promotion instead of being stored.
+//
+// A date set here writes a reminder, which is the one unprompted delivery the
+// persona allows: he asked to be told, on a day he named.
+func promoteCandidate(ctx context.Context, core ipc.CoreAPI, r *http.Request, id int64) (string, error) {
+	doneWhen := strings.TrimSpace(r.FormValue("done_when"))
+	if doneWhen == "" {
+		return "", errors.New("write a definition of done — what has to be true for this to be finished")
+	}
+	text := strings.TrimSpace(r.FormValue("text"))
+	if text == "" {
+		return "", errors.New("empty task text")
+	}
+	due, err := formDue(r, now())
+	if err != nil {
+		return "", err
+	}
+
+	blockedOn, err := resolveBlocker(ctx, core, r.FormValue("blocked_on"))
+	if err != nil {
+		return "", err
+	}
+
+	if err := core.SetTaskFields(ctx, id, doneWhen, blockedOn); err != nil {
+		return "", err
+	}
+	wgt, err := formWeight(r)
+	if err != nil {
+		return "", err
+	}
+	// The importance select is posted whether or not a date is. This ran under
+	// `if due != nil`, so confirming a candidate as "срочно" with no deadline
+	// dropped the word on the floor — the row came back normal and nothing said
+	// why. A promote with neither field set still writes nothing.
+	if due != nil || wgt != 0 {
+		if err := core.EditTask(ctx, id, text, due, wgt); err != nil {
+			return "", err
+		}
+	}
+	if err := core.SetTaskStatus(ctx, id, "open", now(), "tap:web"); err != nil {
+		return "", statusWriteErr(err)
+	}
+	if due == nil {
+		return "confirmed", nil
+	}
+	// A date-only field has no hour. Nine in the morning, because the reminder
+	// is about a day's work and being told at midnight is being told the night
+	// before.
+	fire := time.Date(due.Year(), due.Month(), due.Day(), 9, 0, 0, 0, due.Location())
+	if _, err := core.CreateReminder(ctx, fire, text, ""); err != nil {
+		// The task IS promoted; only the reminder failed. Saying "confirmed"
+		// and nothing else would leave him expecting a nudge that will not come.
+		return "", fmt.Errorf("confirmed, but the reminder did not save: %w", err)
+	}
+	return "confirmed, and maven will remind you that morning", nil
+}
+
+// resolveBlocker turns the blocked-on NAME the form posts into the entity id
+// the row stores. Identity lives in Nexus, so an unresolvable name stops the
+// promotion instead of being written as free text. An empty field is no
+// blocker and reaches Nexus not at all.
+func resolveBlocker(ctx context.Context, core ipc.CoreAPI, field string) (string, error) {
+	name := strings.TrimSpace(field)
+	if name == "" {
+		return "", nil
+	}
+	ref, err := core.ResolveEntity(ctx, name, []string{"person"})
+	switch {
+	case errors.Is(err, ipc.ErrNotImplemented):
+		return "", errors.New("no identity service here, so blocked-on cannot be stored — leave it empty")
+	case errors.Is(err, ipc.ErrNoEntity):
+		return "", fmt.Errorf("nexus does not know %q", name)
+	case err != nil:
+		return "", fmt.Errorf("resolving %q: %w", name, err)
+	case ref.Ambiguous:
+		// Asking, not picking: a task blocked on the wrong person is a
+		// mistake nobody can see afterwards.
+		return "", fmt.Errorf("%q matches %s — say which", name, strings.Join(ref.Candidates, ", "))
+	}
+	return ref.ID, nil
+}
+
+// formWeight reads the importance select. Out-of-range clamps rather than
+// rejects — a bad select is not worth a 400 — but trailing garbage is refused,
+// because strconv is not Sscanf and "3junk" is not a 3.
+func formWeight(r *http.Request) (int, error) {
+	v := r.FormValue("weight")
+	if v == "" {
+		return 0, nil
+	}
+	wgt, err := strconv.Atoi(v)
+	if err != nil || wgt < 0 {
+		return 0, fmt.Errorf("bad weight %q", v)
+	}
+	if wgt > tasks.MaxWeight {
+		wgt = tasks.MaxWeight
+	}
+	return wgt, nil
+}
+
+// formDue reads the date input. An empty field is nil, which on an edit means
+// "clear the date" — the form has no other way to say it.
+func formDue(r *http.Request, now time.Time) (*time.Time, error) {
+	d := r.FormValue("due")
+	if d == "" {
+		return nil, nil
+	}
+	due, err := time.ParseInLocation("2006-01-02", d, now.Location())
+	if err != nil {
+		return nil, fmt.Errorf("bad due date %q", d)
+	}
+	return &due, nil
+}
+
+// fmtTaskDateValue renders a due date the way <input type=date> requires, or
+// "" for no date. Separate from fmtTaskDate, which renders it for reading.
+func fmtTaskDateValue(t *time.Time) string {
+	if t == nil || t.IsZero() {
+		return ""
+	}
+	return t.Local().Format("2006-01-02")
+}
+
+func fmtTaskTime(t *time.Time) string {
+	if t == nil || t.IsZero() {
+		return "—"
+	}
+	return t.Local().Format("02 Jan 15:04")
+}
+
+func fmtTaskDate(t *time.Time) string {
+	if t == nil || t.IsZero() {
+		return "—"
+	}
+	return t.Local().Format("02 Jan")
+}

cmd/mavweb/tasks_test.go

@@ -32,6 +32,31 @@ type fakeTaskCore struct {
 	statusErr error
 
 	promoted bool
+
+	// The promote path's two extra writes.
+	fields   []any
+	edits    []editCall
+	editErr  error
+	fieldErr error
+}
+
+// editCall records one EditTask, so a test can say what the form actually sent
+// down rather than only that the promotion succeeded.
+type editCall struct {
+	ID     int64
+	Text   string
+	Due    *time.Time
+	Weight int
+}
+
+func (f *fakeTaskCore) EditTask(_ context.Context, id int64, text string, due *time.Time, weight int) error {
+	f.edits = append(f.edits, editCall{id, text, due, weight})
+	return f.editErr
+}
+
+func (f *fakeTaskCore) SetTaskFields(_ context.Context, id int64, doneWhen, blockedOn string) error {
+	f.fields = append(f.fields, []any{id, doneWhen, blockedOn})
+	return f.fieldErr
 }
 
 func (f *fakeTaskCore) ListTasks(_ context.Context, status string) ([]ipc.Task, error) {
@@ -220,6 +245,49 @@ func TestApplyTaskPostCarriesWeight(t *testing.T) {
 	}
 }
 
+// Confirming a candidate posts the importance select whether or not a date is
+// set. The weight write hung off `if due != nil`, so "срочно" with no deadline
+// was read off the form and thrown away, and the row came back normal.
+func TestPromoteCandidateCarriesWeightWithoutADueDate(t *testing.T) {
+	core := &fakeTaskCore{}
+	form := url.Values{
+		"action": {"promote"}, "id": {"4"}, "text": {"продлить страховку"},
+		"done_when": {"полис на руках"}, "weight": {"3"},
+	}
+	req := httptest.NewRequest(http.MethodPost, "/tasks", strings.NewReader(form.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	handleTasks(httptest.NewRecorder(), req, core)
+
+	if len(core.edits) != 1 {
+		t.Fatalf("edits = %+v, want the weight written once", core.edits)
+	}
+	if core.edits[0].Weight != 3 || core.edits[0].ID != 4 {
+		t.Errorf("edit = %+v, want id 4 at weight 3", core.edits[0])
+	}
+	if core.edits[0].Due != nil {
+		t.Errorf("edit invented a due date: %v", core.edits[0].Due)
+	}
+	if core.statusVal != "open" {
+		t.Errorf("status = %q, want the candidate promoted", core.statusVal)
+	}
+}
+
+// A promote with neither field set still writes nothing: the row is unchanged
+// apart from its status, and an EditTask here would be a no-op that can fail.
+func TestPromoteCandidateWithNoDateAndNoWeightDoesNotEdit(t *testing.T) {
+	core := &fakeTaskCore{}
+	form := url.Values{
+		"action": {"promote"}, "id": {"4"}, "text": {"продлить страховку"},
+		"done_when": {"полис на руках"}, "weight": {"0"},
+	}
+	req := httptest.NewRequest(http.MethodPost, "/tasks", strings.NewReader(form.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	handleTasks(httptest.NewRecorder(), req, core)
+	if len(core.edits) != 0 {
+		t.Errorf("edits = %+v, want none", core.edits)
+	}
+}
+
 // Out of range clamps rather than 400s; a non-number is a real client error.
 func TestApplyTaskPostClampsWeight(t *testing.T) {
 	core := &fakeTaskCore{created: true}

cmd/mavweb/tools.go

@@ -0,0 +1,112 @@
+package main
+
+import (
+	"cmp"
+	_ "embed"
+	"html/template"
+	"log"
+	"net/http"
+	"strings"
+	"time"
+
+	"github.com/kami/maven/internal/ipc"
+	"github.com/kami/maven/internal/tool"
+	"github.com/kami/maven/internal/webauthn"
+)
+
+//go:embed tools.html
+var toolsHTML string
+
+// toolsTmpl — the enable surface. Server-rendered, no JS: a plain HTML form
+// POSTs back to /tools to enable a proposal. html/template escapes tool names +
+// utterances (they came from voice STT — untrusted text).
+var toolsTmpl = parsePage("tools", toolsHTML, template.FuncMap{
+	"join":       strings.Join,
+	"capability": func(t ipc.Tool) string { return tool.CapabilityOf(t).String() },
+	"risk":       func(t ipc.Tool) string { return string(tool.RiskOf(t)) },
+})
+
+// handleTools serves the enable surface (GET) and applies an enable (POST).
+// POST fields: name, cmd (space-separated argv), destructive (checkbox). cmd is
+// whitespace-split — argv with embedded spaces isn't supported (ponytail: no
+// shell-word parsing; the box owner controls this input, quote a wrapper script
+// if an arg needs spaces).
+func handleTools(w http.ResponseWriter, r *http.Request, core ipc.CoreAPI, session *webauthn.PasskeySession, requireStepUp bool) {
+	if !requireCore(w, core, "tools") {
+		return
+	}
+	ctx := r.Context()
+	var msg string
+	if r.Method == http.MethodPost {
+		if !stepUpGate(w, session, requireStepUp) {
+			return
+		}
+		action := r.FormValue("action")
+		name := strings.TrimSpace(r.FormValue("name"))
+		switch action {
+		case "enable":
+			scope := r.FormValue("scope")
+			cmd := strings.Fields(r.FormValue("cmd"))
+			destructive := r.FormValue("destructive") != ""
+			if name == "" || len(cmd) == 0 {
+				http.Error(w, "name and cmd required", http.StatusBadRequest)
+				return
+			}
+			if err := core.EnableTool(ctx, name, cmd, destructive, scope, time.Now()); err != nil {
+				log.Printf("tools: enable %q: %v", name, err)
+				http.Error(w, "enable failed: "+err.Error(), http.StatusBadGateway)
+				return
+			}
+			msg = "enabled " + name
+		case "disable":
+			if name == "" {
+				http.Error(w, "name required", http.StatusBadRequest)
+				return
+			}
+			if err := core.DisableTool(ctx, name); err != nil {
+				log.Printf("tools: disable %q: %v", name, err)
+				http.Error(w, "disable failed: "+err.Error(), http.StatusBadGateway)
+				return
+			}
+			msg = "disabled " + name
+		case "dismiss":
+			if name == "" {
+				http.Error(w, "name required", http.StatusBadRequest)
+				return
+			}
+			if err := core.DeleteTool(ctx, name); err != nil {
+				log.Printf("tools: dismiss %q: %v", name, err)
+				http.Error(w, "dismiss failed: "+err.Error(), http.StatusBadGateway)
+				return
+			}
+			msg = "dismissed " + name
+		default:
+			http.Error(w, "unknown action", http.StatusBadRequest)
+			return
+		}
+	}
+	proposed, err1 := core.ListTools(ctx, "proposed")
+	enabled, err2 := core.ListTools(ctx, "enabled")
+	if err := cmp.Or(err1, err2); err != nil {
+		log.Printf("tools: %v", err)
+		http.Error(w, "core read failed", http.StatusBadGateway)
+		return
+	}
+	// MCP is off by default and an older core may not know the method at all,
+	// so a failure here renders an empty section rather than breaking the page.
+	servers, err := core.MCPServers(ctx)
+	if err != nil {
+		log.Printf("tools: mcp servers: %v", err)
+		servers = nil
+	}
+	// Enabled rows are shown grouped by capability domain (Vikunja #452). A
+	// flat list stops answering "what can she do to the house" somewhere
+	// around fifteen rows, and that is the question this page exists for.
+	renderPage(w, toolsTmpl, struct {
+		Msg      string
+		Proposed []ipc.Tool
+		Enabled  []ipc.Tool
+		Groups   []tool.CapabilityGroup
+		MCP      []ipc.MCPServerStatus
+	}{msg, proposed, enabled, tool.GroupByDomain(enabled), servers})
+}

cmd/mavweb/trace.html

@@ -1,9 +1,9 @@
 {{template "shellTop" "trace"}}
 <h1>Rule Trace</h1>
-<div class="hint mb-4">{{.Now | ago}} — winner: <strong>{{if .Winner}}{{.Winner}}{{else}}nothing fired{{end}}</strong></div>
+<div class="hint mb-4">{{.Tick.Now | ago}} — winner: <strong>{{if .Tick.Winner}}{{.Tick.Winner}}{{else}}nothing fired{{end}}</strong></div>
 <div class=scroll><table class=mono>
 <tr><th>rule<th>sev<th>predicate<th>gate<th>blocked by<th>detail<th>selected<th>lost to</tr>
-{{range .Rules}}<tr>
+{{range .Tick.Rules}}<tr>
 <td>{{.RuleName}}</td>
 <td>{{.Severity}}</td>
 <td class={{if .PredicateResult}}green{{else}}gray{{end}}>{{.PredicateResult}}</td>
@@ -21,5 +21,25 @@
 <td>{{.LostTo}}</td>
 </tr>{{end}}
 </table></div>
+
+<h1 class=mt-4>Turn Decisions</h1>
+<div class="hint mb-4">Who claimed each utterance, who lost it, and who was never asked. In memory, newest first, cleared on restart.</div>
+{{if not .Turns}}<div class=hint>no turn has run since the daemon started</div>{{end}}
+{{range .Turns}}
+<details class=mb-4>
+<summary><span class=mono>{{.Utterance}}</span> — <strong>{{if .Winner}}{{.Winner}}{{else}}nobody{{end}}</strong> <span class=hint>{{.Ts | ago}}</span></summary>
+<div class=scroll><table class=mono>
+<tr><th>stage<th>claimant<th>would have been<th>score<th>outcome<th>why</tr>
+{{range .Claims}}<tr>
+<td>{{.Stage}}</td>
+<td>{{.Claimant}}</td>
+<td>{{if .Intent}}{{.Intent}}{{else}}—{{end}}</td>
+<td>{{if .HasScore}}{{printf "%.3f" .Score}}{{else}}—{{end}}</td>
+<td class={{if eq .Outcome "won"}}green{{else if eq .Outcome "never_asked"}}red{{else}}gray{{end}}>{{.Outcome}}</td>
+<td>{{.Reason}}</td>
+</tr>{{end}}
+</table></div>
+</details>
+{{end}}
 {{template "shellBottom"}}
 </html>