Bound the wait for a handoff nobody answers

F62. The rotation is agent-driven: the worker asks, and the agent must write
its handoff. When the agent never does, renewals stopped on the ordinary
progress gate, the lease expired, and the task lost an attempt with nothing on
record saying a handoff had ever been requested. Run 16 showed only "agent
status idle and pane unchanged", 34 times.

The request is now stamped, and the wait around it is bounded. While Orchestra
is explicitly waiting the lease renews, because a quiet pane is the answer the
agent was told to give. The request is re-sent once after four minutes, with
the reason it was first asked with. At ten minutes the worker nacks with
failure class handoff_unanswered, and the coordinator releases the task naming
that cause instead of letting the lease die as generic idleness.

The class is known to DebtClassForFailureClass, so a harness that ignores
handoff requests accumulates as its own debt item rather than hiding inside
lease_expired.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CVbaKucEYBjMqVeUgJUsc1
This commit is contained in:
2026-08-29 16:21:35 +04:00
parent 41658aea5b
commit c587f2cc8d
6 changed files with 225 additions and 4 deletions
+20
View File
@@ -2587,3 +2587,23 @@ file.
The ring earned its keep again: one distinct message with a count of 34, rather
than 34 overwrites of one slot.
**Fixed, not yet proven live.** The request is now stamped
(`herdr.Session.HandoffRequestedAt`) and the wait is bounded by
`watchHandoff` in the worker:
- the lease renews while Orchestra is explicitly waiting, because a quiet pane
is the answer the agent was asked for — the renewal gate's new case, bounded
by `handoffAnswerTimeout`;
- the request is re-sent once at `handoffRetryAfter` (4 minutes), with the same
reason it was first asked with;
- at 10 minutes the worker nacks with failure class `handoff_unanswered`, and
the coordinator emits `TaskReleased reason=handoff_unanswered` rather than
letting the lease die as generic idleness.
`DebtClassForFailureClass` knows the class, so a harness that repeatedly
ignores handoff requests now accumulates in the debt ledger instead of hiding
inside `lease_expired`.
Runtime proof still owed: force a request the agent will not answer, and read
the release event rather than the worker state file.