PR Review Summary

Status: Approved

🪜 Strategic-Fit Decision

Per §9 Strategic-Fit Step-Back:

• Decision: Approve
• Rationale: Working PR, zero blocking defect; every close-target AC is delivered and independently verified, and it also clears a live dev regression. Clean Approve — not Approve+Follow-Up (no residual). §9.0 Premise Pre-Flight: no structural-invalidity trigger — the ticket is a well-scoped leaf follow-up with a Contract Ledger + ACs, and the diff matches it.

Peer-Review Opening: Thanks, Emmy — this cleanly closes the four depth-floor edges from my #15143 review, and does it better than the edges asked for: the seed fix derives from the live document instead of swapping one hardcoded snapshot for another. Disclosure: I raised these four edges, so I reviewed each implementation independently against the code rather than assuming my own suggestions were correctly realized. Cross-family review from Claude; CI green at head.

🧭 Patch-Blind Premise Snapshot

• Inputs Read Before Patch: Ticket #15147 (4-row Contract Ledger + 9 ACs); my originating #15143 review (the four edges); current dev DockFlip.mjs (post-#15143/#15146 merge — verified via git merge-base --is-ancestor against the merge commit + a fresh fetch); the changed-file list; the main-tabs Operator seed reality.
• Expected Solution Shape: (1) staging z-index = a safety floor that never lowers a higher captured/effective numeric stacking, exact-restored on cleanup; (2) preserved-landing detection generalized from direct-parent to full ancestor lineage, unchanged lineage still ambiguous; (3) the DockMotion journeys assert semantic prerequisites and derive mutation expectations from the live restored document (not a re-hardcoded snapshot); (4) split/return each independently sampled. Must NOT clone, reparent, override overflow globally, change the motion token, or touch the schema.
• Patch Verdict: Matches / improves. resolveFixedStageZIndex = String(Math.max(2, …numericCaptured/computed)) (captured 7 stays 7, default floors to 2, computed used for the floor but never leaked to inline); captureAncestorLineage + hasAncestorLineageChanged compare the full chain by length + per-index identity (catches an ancestor-only reparent the old check missed); the seed fix derives unrelatedItems/expectedItems from the restored document and proves the insertion index + order — future-proof against the next seed change. The improvement over the ticket: unavailable lineage fail-closes to the bounded wait, and computed-only stacking is covered.
• Premise Coherence: Coheres with friction→gold — non-blocking review edges converted into durable, tested contracts, and a real regression retired, without redesigning the addon. Right-sized: I checked for over-engineering (my edges 1/2/4 were "safe today") and the hardening is proportionate to encoding the missing contracts, not gold-plating.

🕸️ Context & Graph Linking

• Target Issue ID: Resolves #15147
• Related Graph Nodes: #13158 (QT-dock epic), #15137/#15143 (origin of the edges), #15144/#15146 (the standalone-app arc this sits on top of)

🔬 Depth Floor

Documented search (I found no defect, so I'm proving I looked): I actively checked —

1. Ancestor-lineage false-negative: a move that preserves the entire subtree (marker + all ancestors move together) — hasAncestorLineageChanged's length !== + .some((a,i) => a !== first[i]) catches any real cross-boundary move (a genuinely unchanged lineage correctly stays on the bounded ambiguous path). No gap.
2. z-index floor on sub-2 captured values: a pane with captured stacking 0/1 is raised to the 2 safety floor during staging and exact-restored on cleanup — consistent with the AC ("never lowered" is about not dropping a higher value; the floor legitimately raises sub-floor values only transiently). Verified by the "default stacking restores its exact empty inline value" unit assertion.
3. Seed-fix robustness: the DockMotion expectation is derived from the live restored document (unrelatedItems), so it won't re-break when the Operator seed changes again — the correct fix, not a snapshot swap.

One light note (non-blocking): resolveFixedStageZIndex reads getComputedStyle(el).zIndex, which is 'auto' for a position:static marker even when an inline z-index is set — harmless here because the inline value is captured separately and the max() still includes it; just noting the computed leg contributes only for positioned panes.

Rhetorical-Drift Audit: The body's framing ("treats any changed captured ancestor identity as the atomic-landing witness … keeps its safety floor without lowering a higher captured inline or computed numeric z-index … derive their expected order from the restored Operator document") matches the diff exactly. The "Deltas from ticket" (fail-close unavailable lineage; computed-only stacking) are both substantiated (Boolean(firstLineage) && …; the effectiveMarker computed-13 unit test). Findings: Pass.

🧠 Graph Ingestion Notes

• [RETROSPECTIVE]: A clean friction→gold loop closure — a reviewer's four non-blocking depth-floor edges became a scoped follow-up ticket with a Contract Ledger and then durable tests. The standout technique: fixing a stale-seed test regression by deriving expectations from the live restored document (semantic prerequisites + unrelatedItems) rather than swapping one hardcoded whole-document snapshot for another — the snapshot-swap would just re-break on the next seed change.
• [TOOLING_GAP]: The DockMotion/Workstation mounted journeys run nightly/local, not per-PR (known agentos/e2e infra gap); the per-PR safety net here is the 9-test DockFlip unit suite, which does pin the new z-index and ancestor-lineage contracts. Flagged so the merge gate leans on the author's L3 receipts for the runtime ACs.

N/A Audits — 📡 🔗 📑

N/A: no OpenAPI/MCP surface, no skill/convention/AGENTS* change, and no public/consumed-contract change (the body explicitly preserves the motion token and neo.harness.dockZone.v1 schema — the four helpers are addon-internal). Provenance: N/A (hardening of an existing addon; declared anyway).

🎯 Close-Target Audit

• Resolves #15147 (body, newline-isolated) — #15147 labels bug,ai,testing,regression,core — not epic. ✓ Single close-target; no stray Closes/Fixes.

Findings: Pass.

🪜 Evidence Audit

Body declares Evidence: L3 (focused Chromium + Neural Link validation across DockFlip, DockMotion, and standalone Workstation) → L3 required. Achieved ≥ required; no residuals. Includes the seed-regression rerun (--grep addTab 2 passed) — the exact live-red case #15147 called out.

Findings: Pass.

🧪 Test-Evidence & Location Audit

• Execution evidence: exact-head required CI green at eaa22d33 (unit 7m25s, integration-unified, CodeQL, lint×3, lint-pr-body, Analyze, Classify). Author non-CI receipts present and current-head-appropriate: unit 9-passed, DockMotion 7-passed (addTab 2-passed), Workstation 1-passed L3.
• Reviewer falsifier: N/A — no behavior CI leaves unestablished. The new z-index/ancestor-lineage contracts run in the green unit job; the runtime journeys are the author's L3 (nightly-tier), and I traced the source logic directly. I did not re-run the rAF-dependent e2e in the in-app browser (offscreen/rAF-starved would false-negative, per my #15083/#15062 lesson).
• Test location: test/playwright/{unit/dashboard, e2e/dashboard, e2e/workstation} — canonical. ✓

Findings: Pass.

📋 Required Actions

No required actions — eligible for human merge.

📊 Evaluation Metrics

Verdict weights: 30% premise, 30% architecture/placement, 30% diff correctness, 10% AC/audit sanity.

• [ARCH_ALIGNMENT]: 96 — all changes stay inside the DockFlip addon that owns FLIP motion; no token/schema/redesign; the four new methods are cohesive, @protected, and single-purpose. −4: nothing structural, minor cognitive add of a second z-index resolution site (capture + invert).
• [CONTENT_COMPLETENESS]: 96 — @summary JSDoc on every new method; Fat Ticket body with Evidence: line, honest ticket-deltas, provenance.
• [EXECUTION_QUALITY]: 96 — traced z-index max+exact-restore (incl. no-inline-leak), full-chain ancestor detection with fail-close, live-derived seed fix, per-burst sampling; unit suite pins the new contracts; e2e robust. −4 for the mostly-cosmetic dual-resolution redundancy.
• [PRODUCTIVITY]: 97 — all 9 ACs delivered and verifiable, plus a live dev regression retired.
• [IMPACT]: 50 — durable-contract hardening of four motion-boundary edges + one regression fix on a demo/motion addon; healthy but bounded.
• [COMPLEXITY]: 64 — four focused files; the load is the z-index/ancestor logic and the derive-from-live seed reconstruction, not breadth.
• [EFFORT_PROFILE]: Quick Win — high ROI (four contracts hardened + a regression fixed) at contained complexity.

Clean approve, no required actions — eligible for human merge. Nicely done closing the loop, Emmy; the live-document-derived seed fix in particular is the right instinct. 🖖