fix(agents): preserve background completion across parent turns

[rynfar]

Native agents stayed marked Working after completing whenever their parent had moved to another turn. Read-only inspection of the reported conversation confirmed Codex emitted canonical idle/completion events with the original spawn-turn ID and correct session incarnation, but ingestion rejected them at the parent-turn admission gate. Both the persisted Agents roster and in-memory sidebar liveness retained running state.

Treat the four task lifecycle event kinds as session-scoped at the parent admission gates, including while the next parent turn is starting. Preserve provider-instance, session-incarnation, Stop, failed-admission and runtime-current fences. Parent output remains turn-scoped and rejected when stale. No provider-specific protocol or wire changes; the common lifecycle path serves all providers and clients over existing transports.

The sidebar also used the parent's latest session timestamp to time Delegating, so each response restarted its clock. Only Working now shows that timer; the helper refuses settled-parent timestamps. No new UI surface or polling.

Validation:

• 111 focused ingestion tests passed, including completion during ready/running/starting crossed with absent/old admission IDs, both task.updated idle and task.completed, late progress persistence, and stale parent-output rejection.
• The new regressions fail in all six healthy-session cases against the unpatched implementation; all four provider/incarnation/stopped/failed-admission fence tests still pass.
• 239 focused sidebar, subagent-fold, and background-liveness tests passed.
• Server and web typechecks passed; scoped lint/formatting and whitespace checks passed (existing lint warnings remain).
• Independent adversarial review caught the pending-admission race and a malformed test fixture; both resolved, final review found no blockers. No paid-provider trial or live-data writes.

Recovery limitation: this prevents future completion loss. It does not reconstruct already-discarded events in persisted agent history. Restart clears the in-memory sidebar liveness, but old roster entries require fresh authoritative provider status; the patch does not invent completion or time out legitimate long-running agents.

Implemented with GPT-6 Astra in Codex; independent adversarial review used a built-in agent.

---

^{Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
pylon-marketing	Ready	Preview	Sep 17, 2026 5:50pm UTC

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	13.9 KiB	14.0 KiB	+9 B (+0.1%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.2 KiB	7.2 KiB	0 B (0.0%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.7 KiB	6.7 KiB	+9 B (+0.1%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	58.0 KiB	58.0 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	9	9	0 (0.0%)	21	✅
Claude	Total thread wire	13.9 KiB	14.0 KiB	+35 B (+0.2%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.2 KiB	7.2 KiB	+10 B (+0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.7 KiB	6.8 KiB	+25 B (+0.4%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	58.9 KiB	58.9 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	9	9	0 (0.0%)	21	✅

Baseline: 66dc934 · PR result: 2c05062 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 115.7 KiB
• Claude decoded thread snapshot: 116.4 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.