Repository navigation
QA run · ai (5/10) · c91af8ed · 2026-10-17 · 1 PASS / 4 PARTIAL / 0 FAIL / 0 BLOCKED / 5 NOT-RUN #24105
Description
Targeted re-run for the ledger revisions of #24036 (half A), per checklist-test §3: an item whose revision changed a fixture, a step's request or an oracle is re-run once, scoped to what the revision changed. Clauses of those items that the revision did not touch are skipped (out of this run's scope), so an item is pass only where every clause was driven, and partial where the driven clauses passed and the rest were skipped.
Environment
- framework: objectstack
c91af8ed6(origin/main at run start), worktree branchclaude/issue-24036-ledger-a; the branch changes ledger JSON only, no product code - console: objectui
f8267df93aba(this repo's.objectui-sha), bundle built at that sha (check-console-shaOK) - app: showcase under
objectstack dev --ui --seed-admin, posturesingle, file databases in the run's scratch directory. Boot A port 48731 (stock env,OS_PORTexported,OS_LOG_LEVEL=infoafter the first restart). Boot B port 48732 (fresh DB,OS_SEARCH_PINYIN_ENABLED=false,OS_SERVER_TIMING=true,OS_AUTOMATION_SCHEDULED_WORK_ENABLED=1, debug log; laterobjectstack starton the same DB). Port 48733: scratch showcase copy and minimal scratch stacks for build and boot-config legs. Port 48734: a freshcreate-objectstack@nextscaffold - seed: stock showcase seed plus the dev admin; QA personas minted by admin invite plus sign-up (
qa-…-24036@verify.test); area recipesqa-scratch-authz(pinned setsqa_att_reader,qa_att_baseline,qa_att_editor),qa-media-constraints,qa-flow-status-doorsandsearch:qa-contributor-bound-memberas each item names them - window: 2026-10-16 23:08Z to 2026-10-17 01:00Z
Scope
- selector
ai(the bare form ofarea:ai): 10 runnable items pernode scripts/checklist-select.mjs area:ai --json; each row below carries the revision the parser reported - judged 5 of 10; the 5 not-run items are the 4 items this card did not revise plus the revised items whose change was wording only (no re-run owed)
Per-item verdicts
| item | rev | clauses driven | verdict | text oracle evidence |
|---|---|---|---|---|
ai.agent-tool-skill-metadata-roundtrip |
r5 | clause 4 draft read parameter | partial | PUT ?mode=draft 200 state draft; GET ?state=draft 200; GET ?mode=draft 400 VALIDATION_ERROR; bare GET 404 |
ai.mcp-http-surface |
r3 | clause 5 tool list | partial | tools/list returns twelve tools including resume_run |
ai.mcp-run-action-exposure-gate |
r3 | all live clauses | partial | list_actions with objectName -32602, without args [showcase_portfolio_snapshot]; mark_done refused with the exposure prescription; snapshot totals equal REST; sys_* refused |
ai.mcp-validate-expression |
r3 | all clauses on the corrected field | partial | record.estimate_hours * 1.1 ok number; estimate_hourz unknown-field with did-you-mean; type-mismatch warning; context split flow_condition vs formula |
ai.skill-instructions-mcp-prompts |
r2 | clauses 0-4 (runtime and file routes) | pass | prompts capability declared; full listed with title and description; bare and inactive not listed; get returns instructions; bad gets -32602; activation lists it; skill-prompts.test.ts 12 passed |
ai.mcp-stdio-fail-closed |
r6 | none (fixture runner note) | not-run | revised without a re-run owed; os start Origin reading recorded in the item history |
Not run (outside this run's scope, not revised by #24036): 4 items.
Derived verdicts and gaps
- 1 pass, 4 partial, 0 fail, 0 blocked, 5 not-run. No clause failed, so there is no reproduction rule to record and nothing to extract.
- The item edits these runs measured land in the half-A pull request for platform-checklist: revise the items the #23878 full run and the #23879 since:v18 run found stale (knownGaps, steps, fixtures, oracles, source anchors across 15 areas; no clause red) — ledger follow-up #24036; the per-item
historyrows carry the full readings. - Fixture gaps met on the way are recorded in the items' own knownGaps where the revision touched them.
Generated by Claude Code
Activity
objectstack-fleet commentedon Oct 17, 2026
Close-out (
domain:devxseat 2): nofail. Ruling C on #24127 (maintainer): a verification re-run of a ledger edit no longer carries aqa-runrecord, and the existing ones close out now. This was a scoped re-run of the items #24118 revised (50ccfa6697, Part of #24036). Itspartialverdicts mean that clauses the revision did not touch were skipped as out of scope, as this record's table shows; they are not coverage gaps. Itsnot-runitems were unrevised, or revised in wording only.Parked-on: #24036 · ai (5/10) · 4 partial by scope (untouched clauses skipped), named in this record's table; the unmeasured legs the seat record on #24118 lists are carried in the items'
historyrowsGenerated by Claude Code