fix: prevent session stuck at running on turn completion

[sfncore]

What changed

turn.completed events with a mismatched turnId are no longer silently dropped by the STRICT_PROVIDER_LIFECYCLE_GUARD in ProviderRuntimeIngestion. A warning log is emitted instead.

Why

When the runtime emits turn.completed with a turnId that doesn't match the thread's tracked activeTurnId, the guard prevented the thread.session.set command from being dispatched. The session stayed at "running" permanently — the UI showed "working for X minutes" indefinitely with no way to send new messages. Switching threads and back sometimes fixed it (full state resync).

Files changed

• ProviderRuntimeIngestion.ts — turn.completed always applies regardless of turnId mismatch; warning log on mismatch
• ProviderRuntimeIngestion.test.ts — updated 2 tests to match new behavior

All 25 existing tests pass.

---

Note

Medium Risk
Changes provider lifecycle guarding so turn.completed can clear a running session even when turnId mismatches, which could mask real turn-tracking bugs or end a session earlier than expected. Scope is limited to runtime ingestion and covered by updated tests plus a new warning log for observability.

Overview
Prevents threads from getting stuck in a perpetual running state by changing ProviderRuntimeIngestion to always apply turn.completed events even when STRICT_PROVIDER_LIFECYCLE_GUARD detects a turnId mismatch or missing turnId.

When this mismatch occurs under strict guarding, the ingestion layer now emits a warning log and still dispatches the session transition to ready/error. Tests were updated to assert the new behavior (mismatched turn.completed now clears activeTurnId and readies the session).

^{Written by Cursor Bugbot for commit 64ae7ce. This will update automatically on new commits. Configure here.}

Note

Fix session stuck in running state on turn completion when turnId mismatches

• The lifecycle guard in ProviderRuntimeIngestion.ts previously rejected turn.completed events when the turnId didn't match the active turn, leaving sessions stuck in the running state.
• Now turn.completed is always accepted regardless of turnId mismatch or missing active turn, unconditionally transitioning the session to ready.
• A warning is logged with mismatch context (threadId, eventTurnId, activeTurnId) when STRICT_PROVIDER_LIFECYCLE_GUARD is enabled and a mismatch is detected.
• Behavioral Change: a turn.completed from a different turn now clears the running state; previously it had no effect.

^{Macroscope summarized 64ae7ce.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: f378573c-72e1-4ca8-afe3-48341315be34

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[rdoupe]

Adding context from the now-closed PR #1348: that thread explicitly referenced user-filed issues #1069 and #1335 around stale/incorrect Working/Completed status behavior.

Linking those here so #1364 has the same issue lineage in one place.