fix(server): Claude /compact no longer ends early and leaves the thread busy

[t3dotgg]

Problem

Clicking compact on a resumed Claude thread sometimes showed a fake Compacted context 959K → 959K tokens row right away. Then the real compaction ran, and the thread stayed on "Working / Thinking" after it finished.

Root cause, from the provider event log: on resume, Claude first runs a turn of its own to report a background task the previous process left behind. That turn's result (origin: task-notification, num_turns: 0, no prompt echo) arrived 1s after /compact was queued. The adapter took it as the end of the user's turn. So:

1. The /compact turn closed with no compact_boundary. The server added its fallback "compacted" row and guessed the token counts from two noisy usage samples (959,345 → 959,309).
2. The real compaction then ran for 88s with no turn open. Its result (which does echo the turn id) found nothing to close, so the session stayed running.

The same early result also closed normal user turns in 3 other local threads. In those, the reply ended up in a synthetic turn.

Change

Real turns already send their turn id as the prompt uuid. Newer CLIs echo it in user_message_uuids. A result now completes a real turn only if it answers that turn:

• If it echoes prompt uuids, they must include the turn id.
• If it echoes nothing but has a non-human origin, it belongs to a Claude-initiated turn and is skipped. A log line records it.
• Results with neither field (older CLIs) and synthetic turns behave as before.

I checked this rule against all 179 Claude results in my local provider logs. Every skipped result came from a Claude-initiated turn, and every real turn still gets a result that completes it.

Scope and approval

Small, focused fix for a bug Theo reported, in the Claude adapter only. Other providers do not have this result shape.

Verification

• New adapter test replays the recorded order: task-notification result, compact_boundary, then the echoed /compact result. It asserts one turn.completed, after the compaction, and that the compaction event belongs to the turn. It fails without the fix.
• ClaudeAdapter.test.ts and ProviderService.test.ts: 236 passed. Server typecheck and lint for the changed files are clean.

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

🤖 Generated with Claude Code

[coderabbitai]

Navigate logical layers of code changes, visualize relationships, and explore their blast radius.

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository: pingdotgg/t3code/.coderabbit.yaml

Review profile: CHILL

Plan: Team

Run ID: 1db75a97-92e8-4ec8-9692-5fdb885e3adf

📥 Commits

Reviewing files that changed from the base of the PR and between 6b286ae and 7268024.

📒 Files selected for processing (2)

• apps/server/src/provider/Layers/ClaudeAdapter.test.ts
• apps/server/src/provider/Layers/ClaudeAdapter.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 9 remain after this review.

---

📝 Walkthrough

Walkthrough

ClaudeAdapter now skips result messages that belong to another active real turn. A test checks that compaction emits one completion event after the thread-state change and attributes both events to the sent turn.

Changes

Claude turn result handling

Layer / File(s)	Summary
Filter results from other turns apps/server/src/provider/Layers/ClaudeAdapter.ts, apps/server/src/provider/Layers/ClaudeAdapter.test.ts	ClaudeAdapter identifies results for another turn using echoed UUIDs or, when UUIDs are absent, a non-human origin. It skips those results for the active real turn. The test checks event order, count, and turn attribution after /compact.

Priority: ➖ Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Suggested reviewers: juliusmarminge

Merge Risk: 🔵 Low · up to 72680

The fix prevents a compaction turn from closing early. Before merging, confirm that the resolved SDK version declares the new result fields, although the author's reported typecheck suggests it does.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely describes the main fix: preventing Claude /compact from ending prematurely and leaving the thread busy.
Description check	✅ Passed	The description includes complete Problem, Change, Scope and approval, and Verification sections. It explains the root cause, implementation, scope, test coverage, and reported verification results.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 3 functions across 2 files.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Autopilot is currently an internal CodeRabbit preview.

---

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

[macroscopeapp]

Approvability

Verdict: Approved at 7268024

Macroscope's review found this PR approvable — This is a narrowly scoped Claude adapter fix that prevents unrelated result messages from prematurely closing user turns, with a regression test covering the /compact sequence. It introduces no schema, deployment, default, or static-analysis configuration changes.

^{You can add or adjust custom eligibility rules. Learn more.}

[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	13.5 KiB	13.5 KiB	+18 B (+0.1%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.1 KiB	7.1 KiB	−3 B (−0.0%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.4 KiB	6.5 KiB	+21 B (+0.3%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	56.2 KiB	56.3 KiB	+44 B (+0.1%)	66.4 KiB	✅
Codex	Live turn messages	9	10	+1 (+11.1%)	21	✅
Claude	Total thread wire	13.5 KiB	13.5 KiB	−34 B (−0.2%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.1 KiB	7.1 KiB	−4 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.5 KiB	6.4 KiB	−30 B (−0.5%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	57.1 KiB	57.0 KiB	−44 B (−0.1%)	66.4 KiB	✅
Claude	Live turn messages	10	9	−1 (−10.0%)	21	✅

Baseline: 6b286ae · PR result: 7268024 · 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.