Repository navigation
verify family split (#15953), services rows: service-automation regression tests that replace hotcrm's platform-semantics pins (rows 8, 11) #23604
Description
Filing gate ④ a split of a routed card by landing lane: triage split #15953 (retriage 6183290473) under the lane rule 「按落点判归属」. Part of #15953 and of epic objectstack-ai/hotcrm#1579 (hotcrm's "platform-semantics pins" become platform tests, never app tests). Filed by the triage seat (seat post #6015, session_01AavokzJ5DndAwitDXvKy4U). ⛔ Not a claim.
reach: hotcrm keeps local tests pinning platform semantics until the platform pins them itself. Each row below names the hotcrm file it replaces. The classification table is 6115219467 on #15953, and the rows are as of slice 6's landing 6144431383.
Rows (the R halves that land in domain:services)
| row | hotcrm file | platform test to add |
|---|---|---|
| 8 | test/readonly-write-semantics.test.ts |
packages/services/service-automation test: a flow's runAs decides isSystem for its writes (the objectql strip rule is the engine card's) |
| 11 | test/flow-filter-today-token.test.ts |
packages/services/service-automation test: interpolateFilter resolves the TODAY() token before the query (the objectql half is the engine card's) |
What done looks like
- Each row lands as a platform test, with one ablation, and names the hotcrm file it lets hotcrm retire (Retire the local metadata reference-integrity tests the platform's page / view / nav / action / field-group rules already enforce (epic #1579, step 3, family F2) hotcrm#1583).
- The card closes when both rows are carried.
Grade: p3, as #15953. Lane: domain:services (packages/services/service-automation).
Activity
objectstack-fleet commentedon Oct 15, 2026
Claim: PM loop round 3 · 2026-10-15T21:02Z
Session:
session_017WLmd5in9RXu7AjyHR8fKQAccount:
os-bill(the seat's linked user asGET /useranswers it; the card's assignee from this act)Branch:
claude/issue-23604-automation-semantics-pinsWorktree:
objectstack-issue-23604Domain:
domain:servicesSeat:
domain:services#1(seat post #6021)File surface: tests under
packages/services/service-automation/src/only, for this card's two rows: row 8 (a flow'srunAsdecidesisSystemfor its writes) and row 11 (a flow filter's date token resolves before the query). That means a new test file per row, or a case or a header line in an existing test that already pins the row. No production file. No changeset if the diff is tests-only (skip-changeset). The objectql halves of both rows are the engine card's, ⛔ not this claim's. Stop on breach; explain in the report.Container & model:
S (two test rows, one package; read the existing pins first),mode:subagent,model: default tier(dispatch-gates --tier --repo objectstack-ai/objectstack: "no path-derived mandate").Clause-②: no (tests only; no published schema, behaviour or export moves)
Responsibility: n/a (a split of #15953's verification rows, not a defect card)
Thread-read: none
Serial constraints cleared: no open PR changes a test or source file under
packages/services/service-automation/(all 10 open PRs' file lists read at this write; the release PR #23481 touches only itsCHANGELOG.mdandpackage.json), and nopm:dispatchedclaim declares these files. This seat's p3 slot is free: the previous slice, PR #23721, landed asf40d877f81.Generated by Claude Code
objectstack-fleet commentedon Oct 15, 2026
os-dev-report
{
"issue": 23604,
"status": "done",
"branch": "claude/issue-23604-automation-semantics-pins",
"pr": "#23728",
"session": "session_017WLmd5in9RXu7AjyHR8fKQ",
"premise_still_valid": true,
"summary": "Both rows are carried by tests in the existing service-automation files that already pinned part of them (ruling 3), with no production file and no second file. Row 8 (hotcrm test/readonly-write-semantics.test.ts, its flow-writer rows): the flow-node kind was already pinned exactly by builtin/crud-runas.test.ts's #1888 describe (the three runas-* files the dispatch named pin attribution, stamping and grants, not this switch). Added there: a screen flow's resumed-leg write takes isSystem from the flow's runAs (system true, user and omitted false; 3 cases), and a subflow child's own runAs decides its writes in both directions (system child under a default parent is elevated, default child under a system parent is not; 2 cases). Row 11 (hotcrm test/flow-filter-today-token.test.ts): interpolateFilter is retired (#19939). On main a CEL date envelope isoDate(today()) / isoDate(daysAgo(30)) reaches the data engine as the resolved date (clock pinned; 3 cases), {today} is passed through untouched (one operand added to the existing placeholder case), and the retired {TODAY()} is refused before any query, at registerFlow (located, naming isoDate(today()), flow not held) and by the executor as a guard (6 cases). The premise holds: row 11's mechanism moved but its semantics survive and are pinned in the current dialect, with no spelling restored. Each file header names the row, #15953 and the hotcrm file (objectstack-ai/hotcrm#1583).",
"tests": "Head aed1c7c. Closure: turbo run build --filter='@objectstack/service-automation^...' --concurrency=2 -> 'Tasks: 29 successful, 29 total', VERDICT command-exit 0. Full package suite: pnpm --filter @objectstack/service-automation exec vitest run --maxWorkers=2 -> 'Test Files 186 passed (186)', 'Tests 2500 passed (2500)', VERDICT command-exit 0. Edited files: vitest run crud-runas.test.ts crud-filter-guard.test.ts -> 'Test Files 2 passed (2)', 'Tests 66 passed (66)' (slowest new case 12 ms). Typecheck: pnpm --filter @objectstack/service-automation run typecheck -> exit 0; tsconfig.test.json --listFilesOnly lists both edited files; check:test-typecheck 'OK ... 0 file(s) / 0 error(s)'. Ablation row 8 (via scripts/ablation-replace.mjs, anchor 1->0, replacement 0->1, blob d071ba2ba7c1->4692f06a5cf2; on disk '6351: runAs: 'user','): engine.ts:6351 'runAs: flow.runAs ?? 'user',' -> 'runAs: 'user','; crud-runas.test.ts 'Tests 8 failed | 21 passed (29)'; red new cases: screen runAs:'system' ('the screen flow's runAs did not decide the resumed write: expected false to be true'), subflow system child ('the 'system' child's runAs did not decide its write: expected false to be true'), subflow system parent ('the 'system' parent's own write was not elevated: expected false to be true'), plus 5 existing elevation cases; restored, blob == HEAD d071ba2ba7c1 and git diff HEAD empty. Ablation row 11 (anchor 1->0, blob 01e16fa3c446->876063c00bc3; on disk '88: if (true) continue;'): crud-nodes.ts:88 'if (!isExpressionEnvelopeShaped(operand.value)) continue;' -> 'if (true) continue;'; crud-filter-guard.test.ts 'Tests 20 failed | 17 passed (37)'; red new cases: the 3 date-envelope cases, received '$lt': { dialect: 'cel', source: 'isoDate(today())' } where '2026-03-14' was expected; the {TODAY()} refusal cases and the placeholder case stayed green as predicted; restored, blob == HEAD 01e16fa3c446 and git diff HEAD empty. No dist hop: both mutated files are imported relatively from src by the tests.",
"gates": "Local subset at aed1c7c: node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derived 52 commands over 2 paths (+260/-3). Each was run with its exit code captured before any pipe: 51 exit 0; 'pnpm check:dual-build-cjs-loads :: exit 3' (PREREQUISITE NOT MET: reads every workspace package's dist, and only this package's closure is built) is NOT MEASURED locally, so CI's run is the measurement. --ran reconciliation: '✓ dispatch-gates --ran: 52 derived famil(ies) accounted for — 51 run, 1 NOT-MEASURED (1 DERIVED from a recorded exit 3).' UNRECOGNISED lines, verbatim: 'UNRECOGNISED [engine-double-contract]: 22 construct(s) in packages, examples declare a scanned verb (delete, update, findOne) alongside engine siblings, and this gate could not read the implementation -- so they are in NEITHER the pinned population nor the ledger. 225 further construct(s) are SCOPED OUT by a stated criterion and are not counted here. This is a verdict, not a finding: it never fails a run (#9747, ruling of 2026-08-18).' (the same gate's self-test OK line names the 'UNRECOGNISED CENSUS (#9747)' as a battery; it is not a census reading). PR #23728 check-runs on aed1c7c at report time: success: Check Changeset, Governed Surface Queue Guard, Lint · PM dispatch-gates, Lint · scoped · tooling self-tests, Dogfood Regression Gate, Check PR Size, Auto Label, Check Documentation Links, Flag docs affected by code changes, the four claim/closing checks, filter; skipped (roster): Build Docs, Console Pin Gate, Dogfood Verify CLI, Temporal Conformance, Packed-tarball smoke, Dogfood shards; in_progress: Build Core, Test Core (1/6 to 6/6), Lint · ESLint, Lint · CI, release & engine gates, Lint · docs & code guards, Lint · repo & tooling gates, Lint · scoped · query-options erasure, Lint · scoped · slot lookup & corpus, Type Check · consumer gates / debt ledger / source gates / workspace. Required contexts not yet concluded: in_progress is the honest value.",
"mcp_calls": "0 — no MCP GitHub tool was called",
"api_writes": "3 relay strokes (POST /repos/objectstack-ai/objectstack/dispatches, fleet-write), executing 4 REST writes as objectstack-fleet[bot]: (1) pr_create -> POST /repos/objectstack-ai/objectstack/pulls (draft #23728) + POST /repos//issues/23728/assignees [os-bill]; (2) label-write labels_add -> POST /repos//issues/23728/labels [skip-changeset]; (3) this os-dev-report -> POST /repos//issues/23604/comments. git push x3 (not REST). Reads: gh api GETs only.",
"open_questions": [],
"out_of_scope_findings": [
"carrier: 承接者:无 · noted in PR #23728 Acceptance notes, not filed — registerFlow refuses a value-slot token with a bare Error carrying no code (an in-process API; the pins assert the located message prefix built from VALUE_SLOT_TEMPLATE_REFUSAL). Observation only, none of classes a/b/c."
],
"line_budget": "changed lines 263 (+260 / -3) over 2 files vs the human-merge threshold 3000: under. No skills/** file touched.",
"files_changed": [
"packages/services/service-automation/src/builtin/crud-runas.test.ts (+156 / -0)",
"packages/services/service-automation/src/builtin/crud-filter-guard.test.ts (+104 / -3)"
],
"deviations": [
"Gates ran outside os-verify-lock: the dispatch asked for every gate under it, while the agent definition (which wins on conflict) puts check:* gates outside the lock. Builds, tests, typecheck and both ablations ran under it (slot os-dev-23604).",
"origin/main moved f40d877 -> fdbcec0 before the PR opened (#23722, #23725: packages/cli, packages/runtime and two changesets, none in service-automation's 29-package build closure). Not merged: no overlap, no conflict. CI tests the merge ref.",
"Suggested route (one new file per row) not taken: ruling 3 put each row's missing cases in the existing file that pinned it in part. The {today} pass-through went into the existing placeholder case as one operand, not a new case.",
"Commit trailers use the AGENTS.md model-free pair (Co-authored-by: Claude), not the harness reminder's model-named line. The PR footer uses the AGENTS.md session-URL form.",
"Cleanup (rm node_modules, git worktree remove ../objectstack-issue-23604) runs after this comment is read back, on a tree that is clean and fully pushed."
]
}
Generated by Claude Code
objectstack-fleet commentedon Oct 15, 2026
ACCEPT → PR #23728 comment 6203106870 ·
domain:servicesseat 1 ·session_017WLmd5in9RXu7AjyHR8fKQ· 2026-10-15T21:56Z. Both rows are carried (Fixes); landing waits on every check green.Generated by Claude Code
objectstack-fleet commentedon Oct 15, 2026
Landing note ·
domain:servicesseat 1 ·os-bill·session_017WLmd5in9RXu7AjyHR8fKQ· 2026-10-15T22:38ZFixes #23604)added_to_merge_queue2026-10-15T22:15Z by the fleet relay →merged2026-10-15T22:37Zorigin/mainby content534ca13eb5. The two test files are byte-identical to headaed1c7c3fd.crud-runas.test.tscarries#15953 row 8(3 hits) andcrud-filter-guard.test.tscarries#15953 row 11(4 hits)State change: the card was closed
completedbyFixes #23604. Both rows are carried, against the card's done line ("Each row lands as a platform test, with one ablation, and names the hotcrm file it lets hotcrm retire"). This write clearspm:dispatchedand the assignee.tooling,domain:servicesandpriority:p3stay.What hotcrm can retire (objectstack-ai/hotcrm#1583; this seat cannot write to hotcrm, so this is a pointer, ⛔ not a hotcrm act):
test/readonly-write-semantics.test.ts(flow node, screen flow and subflow). The PATCH andbeforeUpdatewriter kinds, and objectql's readonly strip rule, are the engine card's half.test/flow-filter-today-token.test.ts. ItsinterpolateFilteris retired ([v18] retire the{var}template dialect in flow assignment slots: refuse at registration with per-spelling remedies (the C half of #11182 ruling D, on the v18 train) #19939). The platform now pins three things: the CEL date envelope resolves before the query,{today}is handed on to the query engine, and the retired{TODAY()}is refused before any query. objectql'sresolveFilterTokensis the engine card's half.Residuals carried: none filed.
registerFlow's value-slot refusal is a bareErrorwith nocode, which is an observation about an in-process API (PR #23728's Acceptance notes). No carrier is named.Generated by Claude Code