fix(terminal): prevent project action commands from getting stuck

[amitray007]

What Changed

• Prevent project-action commands from getting stuck above the prompt when the terminal has not been opened yet.
• On supported POSIX shells, run the initial action after interactive shell initialization and return to the normal shell when it exits.
• Keep the terminal usable after Ctrl+C stops a long-running initial action.
• Preserve the existing terminal-write fallback for running terminals, Windows, unsupported shells, older servers, and restart paths the server does not handle.

This carries the implementation commit from pingdotgg/t3code#6338 without behavioral changes. Its stable patch ID matches the upstream PR. An Alpha feature record marks the delta as temporary and records the compatibility invariants required during upstream reconciliation.

Why

Opening a fresh terminal and immediately writing a project action can race interactive shell initialization. Commands such as npm run dev may then remain above the prompt without executing. Passing the command during supported shell startup removes that race, while the server acknowledgement keeps every existing fallback path intact.

Related

• Related: fix(terminal): prevent project action commands from getting stuck pingdotgg/t3code#6338
• Related: [Bug]: Project action command gets stuck when the terminal has not been opened pingdotgg/t3code#6337

UI Changes

Before	After

Validation

• vp test run packages/contracts/src/terminal.test.ts apps/server/src/terminal/Manager.test.ts — 80 tests passed
• vp run --filter @t3tools/contracts typecheck
• vp run --filter t3 typecheck — passed with existing Effect suggestions outside the changed terminal code
• vp run --filter @t3tools/web typecheck
• Targeted vp lint and vp fmt --check
• git diff --check
• Stable patch ID matches the complete upstream PR implementation
• The upstream PR records a successful interactive retest of execution, Ctrl+C, and terminal reuse; that runtime pass was not repeated for this Alpha carry

Checklist

• This PR is small and focused
• I explained what changed and why
• I included before/after screenshots for the UI change
• I included a video for the interaction change

Implemented by GPT-5.6 via the Codex harness in T3 Code.

[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	12.6 KiB	12.6 KiB	+12 B (+0.1%)	15.1 KiB	✅
Codex	Thread snapshot wire	6.3 KiB	6.3 KiB	−6 B (−0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.3 KiB	6.3 KiB	+18 B (+0.3%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	51.8 KiB	51.8 KiB	0 B (0.0%)	66.4 KiB	✅
Codex	Live turn messages	16	16	0 (0.0%)	21	✅
Claude	Total thread wire	12.6 KiB	12.6 KiB	−23 B (−0.2%)	15.1 KiB	✅
Claude	Thread snapshot wire	6.3 KiB	6.3 KiB	−7 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.3 KiB	6.3 KiB	−16 B (−0.2%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	52.7 KiB	52.7 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	16	16	0 (0.0%)	21	✅

Baseline: a4cd705 · PR result: 5ece4d8 · 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: 99.9 KiB
• Claude decoded thread snapshot: 100.6 KiB

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