fix(terminal): close completed setup shells safely

[rynfar]

A successfully completed worktree setup script leaves a terminal shell open even when its command has finished. This closes that setup terminal when Pylon has observed its success sentinel and can safely confirm that the launched shell remains at the PTY PID, no child command is running, and no other input or output arrived during inspection. Its output stays in setup history. Failed, reused, active, and uncertain terminals stay open.

Terminal writes and setup cleanup now share the per-thread lock. A write accepted while process inspection is in progress waits until cleanup finishes; if the shell was closed, the write receives the existing terminal lookup error instead of being sent to a process that is about to be killed. A deterministic regression holds inspection open while a write is queued and checks that no extra input reaches the PTY. PTY output arriving during inspection still prevents closure.

This is a partial adaptation of frozen T3 source 1a0c915c41024ddb47e16befd9fc730c1173af02 (#13673). Arbitrary terminals are not closed when a thread settles: a process tree with no children cannot distinguish an idle prompt from an in-process shell command such as read or PowerShell Start-Sleep. That settlement cleanup remains deferred until Pylon has a reliable prompt or command-completion signal. The provider session settlement behavior is unchanged.

Focused verification: terminal manager and setup runner tests passed (94 cases); server t3 typecheck, scoped lint, changed-file format, and diff checks passed. Scoped lint reported one pre-existing warning. No browser or mobile surface change.

Implemented by GPT-6 Codex in the Pylon harness.

[vercel]

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

1 Skipped Deployment

Project	Deployment	Actions	Updated
pylon-marketing	Ignored	Preview	Sep 27, 2026 5:12pm 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	14.0 KiB	14.0 KiB	+33 B (+0.2%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.3 KiB	7.3 KiB	−1 B (−0.0%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.7 KiB	6.8 KiB	+34 B (+0.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	58.0 KiB	58.0 KiB	+44 B (+0.1%)	66.4 KiB	✅
Codex	Live turn messages	9	10	+1 (+11.1%)	21	✅
Claude	Total thread wire	14.0 KiB	14.0 KiB	+39 B (+0.3%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.3 KiB	7.2 KiB	−6 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.7 KiB	6.8 KiB	+45 B (+0.7%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	58.9 KiB	58.9 KiB	+44 B (+0.1%)	66.4 KiB	✅
Claude	Live turn messages	9	10	+1 (+11.1%)	21	✅

Baseline: a247895 · PR result: 1e059cb · 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.