Implement closed-loop workspace follow-ups

infrastructure/inference/AGENTS.md

@@ -15,6 +15,7 @@ Adapter for LLM inference backends. Implements `core:inference` interfaces. No d
 - All LLM responses are proposals — they must be validated by the core before affecting state (invariant #7). Adapters return raw responses; they do not validate.
 - Network calls to inference backends are environment observations; results must be recorded as events by callers if replay must reproduce them (invariant #9).
 - Model lifecycle (spawn/own the llama-server process) lives in `llama_cpp/`; the autonomous scheduler is intentionally unbuilt (see root CLAUDE.md).
+- Model descriptors default to a 24,576-token context window, matching the implementation-stage budget; operators may lower `[[models]].context_size` for constrained hardware.
 - `openai_compat/` is fully remote (no local process/GPU). It speaks `POST {baseUrl}/chat/completions` with `Authorization: Bearer`; `baseUrl` must include the version segment (e.g. `/v1`). It has no `/tokenize`, so it uses a heuristic tokenizer, and no GBNF — JSON artifacts rely on the core's validate-after-retry. Server dispatch keys `[[providers]] type = "nim" | "openai"`; the key comes from `api_key` or `api_key_env`.
 
 ## Work Guidance

infrastructure/inference/commons/src/main/kotlin/com/correx/infrastructure/inference/commons/ModelDescriptor.kt

@@ -10,6 +10,6 @@ data class ModelDescriptor(
     val modelPath: String,
     val residencyMode: ResidencyMode,
     val idleTimeoutMs: Long = 60_000L,
-    val contextSize: Int = 8192,
+    val contextSize: Int = 24_576,
     val capabilities: Set<CapabilityScore> = emptySet(),
 )