[lukemaj]

What: A child thread's header reads project / parent title / thread title again; the parent title opens the parent, and a small menu switches between sibling threads.
Why: The V2 port dropped v1's header crumbs (#18); V2 only shows "Subagent of" at the top of the conversation, which scrolls away.
So what: The user approved merging this; it ships in the next Chromeria build.

Closes #201.

How

Ported v1's ThreadParentCrumbs onto V2's thread lineage (lineage.parentThreadId, relationship subagent; forks excluded). Pure helpers in threadParentCrumbs.logic.ts; one 3-line hook in upstream ChatHeader.tsx (already carried by fork features). Only child threads subscribe to the full thread-shell list, so top-level threads pay nothing.

Proof

• vp test run apps/web/src/components/chat/threadParentCrumbs.logic.test.ts: 3 passed.
• tsc --noEmit for apps/web: passed. vp lint/vp fmt on changed files: no errors.
• No screenshots: no browser started without permission; the user verifies in the next build.

Elon record

Requirements and who asked: User asked to bring back the stacked parent title for child threads.
Deleted: Grandparent chains, mobile, the v1 Agents panel tree.
Bottleneck: None; V2 thread shells already carry lineage.
Checked myself: v1 source at ffcdae87ca, V2 lineage writers (SubagentProjection, lineageBackfill, ThreadForkService), test and typecheck output.

Model: Claude Opus 5.5 in Claude Code (Chromeria thread).

🤖 Generated with Claude Code

[lukemaj]

Independent review (GPT-6 Luna, source only) on a64fdeff: no blocking findings. Child threads from childThreads, Prism delegate_task, Spectrum and provider subagents carry subagent lineage on EnvironmentThreadShell; forks excluded; hook order stable; shell-list subscription gated on a resolved parent; long titles truncate.

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ The exact PR base did not have a successful artifact. Baseline uses the latest successful main measurement shown below.

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	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.9 KiB	20.9 KiB	0 B (0.0%)	29.3 KiB	✅
Codex	Live turn messages	2	2	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	−7 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	1.2 KiB	1.2 KiB	−7 B (−0.6%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	0 B (0.0%)	29.3 KiB	✅
Claude	Live turn messages	2	2	0 (0.0%)	8	✅

Baseline: 31aa793 · PR result: 98be8cd · 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.