[tomcorke]

Closes #7

Summary

Task documents were flat and hard to navigate by their own metadata, and there was no UI to repair a moved or changed-ID file. This adds pinned and state groups, draft title/state/pin editing, explicit relink reconciliation, and clearer task-thread context.

Changed areas

• The task sidebar groups pinned tasks first, then groups the rest by state, with a dedicated No state group; task-linked threads show their project, provider, and runtime status.
• Task properties are patched into existing YAML without rewriting comments, unknown fields, or the Markdown body. Unsafe frontmatter is rejected instead of producing malformed YAML.
• Relink controls recover missing or changed-ID files, with explicit confirmation for adopting a new ID and server-side duplicate checks.
• The saved-usage retention test now freezes its service clock so its fixed transcript fixture stays inside the 92-day retention window.
• User task guidance covers grouping, property edits, and relinking.

Evidence

Before:

After:

Validation: vp run check:changed reported every stage passing: format, lint, GraphQL, typecheck across 32 packages, 66 tests across 5 files (2 skipped), knip, and React Compiler. The focused retention test also passes with the frozen clock.

Merge danger

Door: task sidebar navigation and the task document pane, plus the existing task relink RPC.

Blast radius: task-document metadata parsing, save/relink behavior, and task-linked thread context in the shared web/desktop UI. No provider protocol or orchestration changes.

Model/harness: Codex gpt-6-luna, T3 Code, max reasoning effort with the default service tier.

[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	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.3 KiB	1.3 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	21.6 KiB	21.6 KiB	0 B (0.0%)	24.4 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.1 KiB	5.0 KiB	−6 B (−0.1%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.3 KiB	1.3 KiB	−6 B (−0.5%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 9c3ce12 · PR result: fa5fe06 · 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: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

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