feat(server): use the selected GitHub login in terminals and agent sessions

[donnes]

The GitHub CLI account chosen in Settings → Source Control only applied to T3 Code's own gh commands. Terminals and agent sessions used gh's globally active login, so an agent running gh pr create could act as the wrong account. Switching with gh auth switch elsewhere also changed it mid-task, because that rewrites one config file shared by every process.

Fix

• GitHubCliAccountEnvironment in sourceControl/GitHubCliAccountSelection.ts resolves the checkout's selected login, reads its token with gh auth token --user (cached for 1 minute), and returns it as GH_TOKEN/GITHUB_TOKEN, or the enterprise variables for GHES hosts. The active login is never switched, so multiple terminals, agents and T3 instances can use different accounts at once.
• Terminals merge it into the shell environment at spawn.
• Every provider adapter merges it into its per-session process environment: Claude, Codex, Cursor, Grok, OpenCode and Antigravity. An OpenCode server reached by serverUrl keeps its own environment, because T3 doesn't start it.
• If the selected login can't be read, the process starts with gh's active login and a warning is logged. T3's own GitHub actions still fail closed.
• User docs updated. They previously said terminals were unaffected.

Changing the selection applies to terminals and sessions started afterwards.

Tests

• New: token-lookup behavior (github.com vs enterprise hosts, signed-out fallback, caching).
• New: terminal spawn env.
• Existing adapter, terminal and source-control suites pass.
• AntigravityAdapter.test.ts › serves client file reads and writes only inside the session roots fails identically on main on macOS (a /var vs /private/var temp-path mismatch) and is unrelated.

Done with Claude Opus 5.5 in Claude Code (T3 Code).

🤖 Generated with Claude 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	13.6 KiB	13.5 KiB	−58 B (−0.4%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.1 KiB	7.1 KiB	−10 B (−0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.5 KiB	6.4 KiB	−48 B (−0.7%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	56.3 KiB	56.2 KiB	−44 B (−0.1%)	66.4 KiB	✅
Codex	Live turn messages	10	9	−1 (−10.0%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	+17 B (+0.1%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	+9 B (+0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.4 KiB	6.4 KiB	+8 B (+0.1%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.0 KiB	57.0 KiB	0 B (0.0%)	66.4 KiB	✅
Claude	Live turn messages	9	9	0 (0.0%)	21	✅

Baseline: 1c70739 · PR result: 9ffba12 · 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: 114.0 KiB
• Claude decoded thread snapshot: 114.7 KiB

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

[donnes]

Reviewed all 15 changed files at 6b350b7. No verified correctness bugs or blocking findings. Two nonblocking test suggestions are attached inline.

Validation: 512 tests passed across nine focused suites covering account selection, terminal management, all six provider adapters, and GitHub CLI behavior. Server typecheck and diff whitespace checks passed.

Reviewed with GPT-6-Astra through the Codex harness in T3 Code.