Weekly periodic evals: red lane needs triage

Automated weekly report — run: https://github.com/garrytan/gstack/actions/runs/33445027170

- periodic reconciliation exit: 1
- gate census job: success

```
[test:paid] report: 6/6 slices, 71 planned shards, tier=periodic
slice 1 timed-out 1800s test/carve-section-loading.test.ts
slice 1 passed 19s test/llm-judge-recommendation.test.ts
slice 1 failed 26s test/skill-e2e-autoplan-chain.test.ts
slice 1 passed 0s test/skill-e2e-coverage-audit.test.ts
slice 1 passed 0s test/skill-e2e-gbrain-roundtrip-local.test.ts
slice 1 passed 181s test/skill-e2e-office-hours-brain-writeback.test.ts
slice 1 failed 721s test/skill-e2e-plan-ceo-mode-routing.test.ts
slice 1 failed 1502s test/skill-e2e-plan-devex-finding-count.test.ts
slice 1 passed 243s test/skill-e2e-plan-prosons.test.ts
slice 1 passed 114s test/skill-e2e-retro.test.ts
slice 1 failed 350s test/skill-e2e-setup-gbrain-path4-local-pglite.test.ts
slice 1 passed 0s test/skill-e2e-triage.test.ts
slice 2 passed 0s test/codex-e2e-plan-format.test.ts
slice 2 passed 52s test/skill-e2e-arm-benchmark.test.ts
slice 2 passed 600s test/skill-e2e-autoplan-dual-voice.test.ts
slice 2 passed 327s test/skill-e2e-cso.test.ts
slice 2 passed 0s test/skill-e2e-hermetic-canary.test.ts
slice 2 passed 41s test/skill-e2e-office-hours-phase4.test.ts
slice 2 failed 600s test/skill-e2e-plan-ceo-review-section-loading.test.ts
slice 2 failed 736s test/skill-e2e-plan-eng-finding-count.test.ts
slice 2 passed 0s test/skill-e2e-plan-tune-cathedral.test.ts
slice 2 passed 218s test/skill-e2e-review-army.test.ts
slice 2 passed 61s test/skill-e2e-setup-gbrain-remote.test.ts
slice 2 passed 0s test/skill-e2e-workflow.test.ts
slice 3 passed 0s test/codex-e2e-recommendation-substance.test.ts
slice 3 passed 255s test/skill-e2e-auq-consistency.test.ts
slice 3 passed 8s test/skill-e2e-benchmark-providers.test.ts
slice 3 passed 0s test/skill-e2e-deploy.test.ts
slice 3 passed 0s test/skill-e2e-ios-device.test.ts
slice 3 passed 121s test/skill-e2e-office-hours.test.ts
slice 3 failed 521s test/skill-e2e-plan-ceo-split-overflow.test.ts
slice 3 failed 1226s test/skill-e2e-plan-eng-finding-floor.test.ts
slice 3 passed 0s test/skill-e2e-plan-tune.test.ts
slice 3 passed 0s test/skill-e2e-review-attribution.test.ts
slice 3 passed 194s test/skill-e2e-ship-section-loading.test.ts
slice 3 passed 0s test/skill-llm-eval-spec.test.ts
slice 4 passed 0s test/codex-e2e-sol-scope.test.ts
slice 4 passed 444s test/skill-e2e-auq-matrix.test.ts
slice 4 passed 0s test/skill-e2e-bws.test.ts
slice 4 passed 539s test/skill-e2e-design.test.ts
slice 4 passed 0s test/skill-e2e-ios-swift-build.test.ts
slice 4 passed 16s test/skill-e2e-opus-47.test.ts
slice 4 passed 656s test/skill-e2e-plan-design-finding-count.test.ts
slice 4 failed 1004s test/skill-e2e-plan-eng-multi-finding-batching.test.ts
slice 4 passed 994s test/skill-e2e-plan.test.ts
slice 4 passed 117s test/skill-e2e-review.test.ts
slice 4 passed 0s test/skill-e2e-skillify.test.ts
slice 4 passed 119s test/skill-llm-eval.test.ts
slice 5 passed 0s test/codex-e2e.test.ts
slice 5 passed 158s test/skill-e2e-auq-verbose-vs-carved-ab.test.ts
slice 5 passed 147s test/skill-e2e-conductor-prose.test.ts
slice 5 passed 75s test/skill-e2e-diagram.test.ts
slice 5 passed 0s test/skill-e2e-learnings.test.ts
slice 5 failed 1663s test/skill-e2e-overlay-harness.test.ts
slice 5 passed 48s test/skill-e2e-plan-design-finding-floor.test.ts
slice 5 passed 175s test/skill-e2e-plan-eng-plan-mode.test.ts
slice 5 passed 277s test/skill-e2e-qa-bugs.test.ts
slice 5 passed 0s test/skill-e2e-session-intelligence.test.ts
slice 5 passed 0s test/skill-e2e-spec-execute.test.ts
slice 5 failed 136s test/skill-routing-e2e.test.ts
slice 6 passed 0s test/gemini-e2e.test.ts
slice 6 failed 284s test/skill-e2e-auto-decide-preserved.test.ts
slice 6 passed 108s test/skill-e2e-context-skills.test.ts
slice 6 passed 9s test/skill-e2e-first-task-scaffold.test.ts
slice 6 passed 0s test/skill-e2e-memory-pipeline.test.ts
slice 6 failed 890s test/skill-e2e-plan-ceo-finding-count.test.ts
slice 6 passed 157s test/skill-e2e-plan-design-plan-mode.test.ts
slice 6 passed 301s test/skill-e2e-plan-format.test.ts
slice 6 passed 171s test/skill-e2e-qa-workflow.test.ts
slice 6 passed 76s test/skill-e2e-setup-gbrain-bad-token.test.ts
slice 6 passed 0s test/skill-e2e-third-party-actions.test.ts
```

Exclusion policy: test/helpers/periodic-exclude-data.ts (every entry needs reason + tracking; removal re-activates the file next week).

Start with the linked weekly Actions run and reproduce the failed or timed-out tests listed in the report, including test/carve-section-loading.test.ts and the failing skill-e2e suites. Read test/helpers/periodic-exclude-data.ts before considering an exclusion. Done means the red tests are repaired or each exclusion has a reason and tracking information.

Written by the indexing model from the issue text.