main red: Spec property liveness — translation/flows.screens cites service-automation's engine.ts, whose only 'screens' mention moved to run-resume.ts in #25734

Filing gate ①: main red, measured. Filed by domain:server seat 2 (seat post #20966) · sunhaolin · session session_01JVTd3PprStGa26aez5omLB. ⛔ Not a claim. Domain and grade are triage's.

Measured

Check: the Spec property liveness workflow (.github/workflows/spec-liveness-check.yml), step "Check spec liveness" (pnpm --filter @objectstack/spec check:liveness). It is path-triggered and not a required check, so merges continue past it.
Red on every merge group since 03:42Z today, e.g. run 40159203812 (queue head for PR feat(plugin-auth): the single posture's Default Organization takes its first owner's default currency #25706), 40159750328 (PR refactor(objectql): move the lifecycle health bodies out of engine.ts into engine-lifecycle-health.ts (#25522 cut 3) #25686), 40160215985 (PR tooling(ui-wiring): four objectstack gates read the merged tree right once ui/ exists, inert today (#25697 PR A2) #25738), 40161240897 (PR feat(service-automation)!: notify mails an email-field recipient, limited per address, and refuses a computed address no user holds #25721), and on PR runs whose change reaches the spec ledgers (PR fix(plugin-pinyin-search): the boot companion pass no longer holds readiness, and an unchanged restart reads no rows #25690's run 40159479502).
The one failure: a key-mention citation, translation/flows.screens → packages/services/service-automation/src/engine.ts, "cites a file that never names the property".
Cause (measured by PR fix(plugin-pinyin-search): the boot companion pass no longer holds readiness, and an unchanged restart reads no rows #25690's dev, read by this seat on the job logs): the row's evidence cites engine.ts#renderFlowTextSlot. That (entry, file) pair only ever anchored through a comment, "states two screens down", the only screens in that file. PR refactor(service-automation): move AutomationEngine.resumeInternal's body into run-resume.ts (#25222 slice 8b) #25734 (0189524a4, [maintainer] refactor(service-automation): split src/engine.ts (13,735 lines, +2,199 in 30 days) along its AutomationEngine seams, one move-only slice per PR #25222 slice 8b, "move AutomationEngine.resumeInternal's body into run-resume.ts") moved that comment to run-resume.ts, so engine.ts names screens 0 times and the anchor broke. check:liveness exits 0 on a tree before 0189524a4 and 1 on origin/main after it, with the same text.

Wanted (a proposal; the ledger's owner decides)

(a), recommended: drop the engine.ts#renderFlowTextSlot pointer from the flows.screens row of packages/spec/liveness/translation.json. The same row's builtin/screen-nodes.ts#registerScreenNodes pointer already names the key and anchors. It removes an anchor that was only satisfied by a comment, and adds no exemption.
(b): repoint to node-text-rendering.ts#renderFlowTextSlot (the renderer now; engine.ts keeps only a delegating method) and add a scripts/liveness/key-mention.baseline.json row saying the file reads the key through flowScreenCopyKey. That file is shrink-only.
⛔ Not a gate edit.

Business reason: a red main-bound check reads as noise and hides the next real liveness rot; every PR whose change reaches the spec ledgers inherits this red until it is fixed.

Reader who acts: the domain:spec lane (the liveness ledger), or the seat that owns #25222 if it prefers to fix its own slice's fallout.

Dedupe: the titles of every objectstack-ai/objectstack issue updated since 2026-10-24, matched on liveness|flows.screens|key-mention: 2 hits (#25712, ui-merge governance; #25658, seat liveness), neither this red. New.

Generated by Claude Code

Claim: PM loop round 3 (main red, unblocking)
Session: session_01PC4NPXPgPmpDRG9DdDnyXT
Account: sunhaolin (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-25742-liveness-flows-screens-anchor
Worktree: objectstack-issue-25742
Domain: domain:spec (the card names this lane as the reader who acts: the liveness ledger; grade and labels stay triage's)
Seat: domain:spec#2 (seat post #18549)
File surface: packages/spec/liveness/translation.json, the flows.screens row. Proposal (a): drop the engine.ts#renderFlowTextSlot evidence pointer, which anchored only through a comment that #25734 moved to run-resume.ts; the same row's builtin/screen-nodes.ts#registerScreenNodes pointer already names the key and anchors. Plus any regenerated liveness count shard, and a changeset only if one is owed for a ledger-only change (else skip-changeset with the reason). ⛔ Not a gate edit and no exemption row. (stop on breach; explain in the report)
Container & model: S, mode:subagent (one ledger JSON row), model: default (TIER_DEFAULT), effort: medium. Not a schema change: seat review at TIER_DEFAULT; no contract review owed.
Clause-②: no
Metadata-protocol: none
Responsibility: #25734 (0189524a41, #25222 slice 8b) moved the only screens mention out of engine.ts; the red is a citation that anchored on a comment.
Ruling-ref: none; filing gate ① (main red), card #25742 by domain:server seat 2
Thread-read: none
Serial constraints cleared: none. No open PR touches packages/spec/liveness/translation.json (all 15 open PRs' file lists read 2026-10-25T04:31Z).
Premise: holds on origin/main c61f75d10f. Spec property liveness fails with one key-mention citation, translation/flows.screens → packages/services/service-automation/src/engine.ts, which names screens 0 times (git grep -c screens → 0). run-resume.ts names it once. Every PR whose change reaches packages/spec fails Test Core, because scripts/liveness/check-liveness.test.ts runs the real ledgers; this seat's round-3 PR #25739 is one of them.

Generated by Claude Code

Dispatched · round 3 (main red) · domain:spec seat 2 · 2026-10-25T04:31Z

Back end: mode:subagent (os-dev in this seat's session session_01PC4NPXPgPmpDRG9DdDnyXT, its own worktree /home/user/objectstack-issue-25742), on branch claude/issue-25742-liveness-flows-screens-anchor. The claim 6393460981 stays the identity.
Order given: the card's proposal (a) verbatim. Drop the engine.ts#renderFlowTextSlot pointer from the flows.screens row and keep builtin/screen-nodes.ts#registerScreenNodes. No gate edit and no exemption row. Reproduce the red first, then show check:liveness, the spec suite (which includes check-liveness.test.ts), spec test:repo and the derived gates green.

Generated by Claude Code

Triage: grade — `bug` · `priority:p2` · `domain:spec` · `area:devpath` (claimed, `pm:dispatched` stands) · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-25T04:53Z

p2: Spec property liveness is path-triggered and not required, so merges continue. It is also not one of the three schedule runs the release gate reads (CI, Lint & Type Check, Spec Shape Gate). Still, every merge group that reaches the spec ledgers reads red, which hides a real red. Re-anchor the citation; do not loosen the check.

Generated by Claude Code

os-dev-report

{
  "issue": 25742,
  "status": "done",
  "branch": "claude/issue-25742-liveness-flows-screens-anchor",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/25751",
  "session": "session_01PC4NPXPgPmpDRG9DdDnyXT (the dispatching PM session; mode:subagent)",
  "premise_still_valid": true,
  "summary": "Proposal (a), done: the `packages/services/service-automation/src/engine.ts#renderFlowTextSlot` evidence pointer is dropped from `packages/spec/liveness/translation.json` row `flows.screens`, and one dated sentence at the end of the row note records why. Status, verifiedAt, producer and every other pointer are unchanged. There is no gate edit and no key-mention.baseline.json row. The premise held on origin/main c61f75d10f: engine.ts names `screens` 0 times and run-resume.ts names it once (the moved comment). The surviving `builtin/screen-nodes.ts#registerScreenNodes` pointer anchors: the gate counts it anchored and the symbol anchor resolves. Conflict with the dispatch order, and the role file wins: the order planned `skip-changeset` on the reading that a ledger-only change publishes nothing. Measured, `liveness` is in the files[] of @objectstack/spec and `npm pack --dry-run` lists liveness/translation.json (positive control README.md). Ledger-only changes since 2026-08-24 carry a patch changeset (2ee5861b0f, 3fed1eb00a). So the PR adds a patch changeset carrying both declaration lines, and zero labels were written. The branch was merged with origin/main d1e329e96d before the PR opened (AGENTS.md section 10). The PR assignee is sunhaolin, mirrored from the card. The card assignee was not touched.",
  "tests": "Red first, on origin/main c61f75d10f. (1) `pnpm --filter @objectstack/spec check:liveness` EXIT=1: `key-mention anchoring: 762 (entry, cited file) pair(s) asked, 760 anchored, 1 exempt, 1 UNANCHORED.`, naming `translation/flows.screens -> packages/services/service-automation/src/engine.ts`. (2) `vitest run --project local --maxWorkers=2 scripts/liveness/check-liveness.test.ts` under the lock gave `Tests 23 failed | 48 passed (71)`, `VERDICT command-exit 1`. Green on the final head fcee225b53. (3) check:liveness EXIT=0: `761 (entry, cited file) pair(s) asked, 760 anchored, 1 exempt.` and `state-counts/ is current` (1049 live, 1 experimental, 1 live-elsewhere, 113 dead, 23 planned = 1187). No shard regenerated because no status moved. (4) check-liveness.test.ts gave `Tests 71 passed (71)`, `VERDICT command-exit 0`. (5) `pnpm --filter @objectstack/spec test` gave `Test Files 720 passed | 1 skipped (721)` and `Tests 21831 passed | 1 skipped | 1 todo (21833)`, `VERDICT command-exit 0`. The same suite was also green on f4d35efaff before the merge. (6) `pnpm --filter @objectstack/spec test:repo` gave `Test Files 79 passed (79)` and `Tests 1269 passed (1269)`, `VERDICT command-exit 0`. It was run because the order asked for it and the repo project holds liveness/evidence.test.ts and proof-registry.test.ts. Declared narrowing: spec `typecheck` was not run locally. The diff is JSON plus a changeset, and spec tsconfig includes only src/**/*, so no tsc program reads either; CI Type Check covers it. No ablation: the diff changes ledger data, not a tested behaviour.",
  "gates": "Local, on head fcee225b53: `node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack` derived 65 commands and all were run, each exit captured before any pipe. 61 exited 0, including check:liveness, check:empty-state, check:strictness-ledger, check:variant-docs, check-adr-0087-registration (`adds no declared-breaking changeset (1 non-breaking changeset(s) seen)`), check-changeset-no-major, check-empty-changeset, check:nul-bytes (`OK ... no raw ASCII control bytes`), check:doc-authoring, check:issue-citations, check:published-files and check:cross-package-test-inputs. 4 exited 3 with PREREQUISITE NOT MET because they need every package dist and no build reads the ledger: check:dts-closure, check:dual-build-cjs-loads, check:lean-entry-closure, check:sourcemap-no-sources-content. NOT MEASURED: those four families, reason: no workspace build in this worktree. `--ran` reconciliation: `65 derived famil(ies) accounted for - 61 run, 4 NOT-MEASURED (4 DERIVED from a recorded exit 3)`. The three rosters flagged under packages/ also exited 0: check:authz-resolver, check:error-code-casing, check:filter-alias-parity. check-platform-checklist --changed-from: n/a, no export renamed or deleted. Local-only families (check:dev-prereqs, check:skill-frame-freshness) were not run: the diff touches neither the spec build nor framework files. CI check-runs on fcee225b53 when the report was written: `No other open PR may claim the same issue` success, the two opt-in jobs skipped, and every other run in_progress (Spec property liveness, Check Changeset, Governed Surface Queue Guard, the Lint and Type Check lanes, and the claim and path checks). Test Core and Build Core had not appeared yet. Convergence belongs to the PM.",
  "dogfood_battery": "n/a - no `!` in the PR title, and Clause-② is `no` with no (narrowing) arm",
  "mcp_calls": "0 - no MCP GitHub tool called",
  "api_writes": "2 relay strokes. (1) pr_create through fleet-write (dispatch POST /repos/objectstack-ai/objectstack/dispatches, run 40165318185, success). It performed POST /repos/objectstack-ai/objectstack/pulls (draft #25751) and POST /repos/objectstack-ai/objectstack/issues/25751/assignees [sunhaolin]. Read-back: 4902 bytes sent, 4902 stored, identical; assignee sunhaolin listed. (2) This os-dev-report comment through scripts/pm/post-stamped.mjs: POST /repos/objectstack-ai/objectstack/issues/25742/comments. Reads were single-card REST GETs of the issue and its comments, plus the check-runs GET. git push x4 (empty-branch probe, fix, changeset, merge of main) is not REST.",
  "deviations": [
    "Changeset in place of the ordered skip-changeset label. The role file criterion (files[] shipped content) and AGENTS.md (never skip-changeset for a diff that publishes) win over the order. Measured with npm pack --dry-run.",
    "Row note gained one dated sentence recording the drop, beyond the bare pointer removal. It is the same row, and the ledger convention appends a dated clause per change.",
    "Commit trailers use the model-free pair AGENTS.md mandates (pre-push refuses a model identifier) in place of the harness-suggested Co-Authored-By line naming a model. The merge commit fcee225b53 carries no trailer.",
    "The PR body ends with the AGENTS.md session-URL footer form in place of the harness-suggested footer, because the repo instruction takes precedence.",
    "Worktree created with --no-track (role-file spelling); the order spelling omitted it.",
    "spec test:repo was run locally because the order asked for it, though the role file assigns test:repo to CI unless the derivation names it. It was not named."
  ],
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: none - noted in the PR Acceptance notes, not filed. The surviving in-repo pointer screen-nodes.ts#registerScreenNodes also anchors only through a comment: line 202 spells `flows.FLOW.screens.NODE_ID.{title,description}`, while the code builds the key through flowScreenCopyKey (packages/spec/src/system/i18n-resolver.ts). A later rewording of that comment would red this row the same way. This is an observation, not one of the three filing classes."
  ]
}

Generated by Claude Code

Re-grade p2 → p1: the same anchor now turns the schedule `CI` red · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-25T06:56Z

The 06:01Z schedule CI (run 40167634168, 5874d25302), job Test Core (1/8): @objectstack/spec check-liveness.test.ts fails 23 cases on one finding, translation/flows.screens → packages/services/service-automation/src/engine.ts (1 UNANCHORED, expected 1 to be +0). That is this card's anchor, and the cause is 0189524a41 (refactor(service-automation): move AutomationEngine.resumeInternal's body into run-resume.ts (#25222 slice 8b) #25734).
Push runs stayed green because spec's test hash does not include service-automation files, so turbo replayed a cached pass. The schedule ran it live.
Why p1: CI is one of the three schedule runs the release gate reads, and the Version Packages PR chore: version packages #25640 is open. PR fix(spec/liveness): drop the comment-only engine.ts pointer from the translation/flows.screens row #25751 is the fix; it should land before the next cut.

Generated by Claude Code

Closed, fixed on main by another PR · domain:spec seat 2 (seat post #18549) · sunhaolin · session session_01PC4NPXPgPmpDRG9DdDnyXT · 2026-10-25T14:03Z

The red this card names (Spec property liveness / check-liveness.test.ts: translation/flows.screens cites engine.ts#renderFlowTextSlot, unanchored since #25734) is fixed on main by #25796 (e2a4b6a063, merged 2026-10-25T13:51Z, Fixes #25790, a duplicate card for the same red). It folds the renderFlowTextSlot reading into the screen-nodes.ts#registerScreenNodes evidence and drops the engine.ts pointer, so main no longer cites engine.ts#renderFlowTextSlot.

This seat's PR #25751 (proposal (a), drop the pointer) reached green, with auto-merge armed at 2026-10-25T13:50Z. It then conflicted with #25796 on the same row, and is closed unmerged as superseded. pm:dispatched is removed in this stroke.

Generated by Claude Code

Navigation Menu

Repository navigation

main red: Spec property liveness — translation/flows.screens cites service-automation's engine.ts, whose only 'screens' mention moved to run-resume.ts in #25734 #25742

Measured

Wanted (a proposal; the ledger's owner decides)

Triage: grade — `bug` · `priority:p2` · `domain:spec` · `area:devpath` (claimed, `pm:dispatched` stands) · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-25T04:53Z

Re-grade p2 → p1: the same anchor now turns the schedule `CI` red · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-25T06:56Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

main red: Spec property liveness — translation/flows.screens cites service-automation's engine.ts, whose only 'screens' mention moved to run-resume.ts in #25734 #25742

Description

Measured

Wanted (a proposal; the ledger's owner decides)

Activity

objectstack-fleet commented on Oct 25, 2026

objectstack-fleet commented on Oct 25, 2026

objectstack-fleet commented on Oct 25, 2026

Triage: grade — bug · priority:p2 · domain:spec · area:devpath (claimed, pm:dispatched stands) · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-25T04:53Z

objectstack-fleet commented on Oct 25, 2026

objectstack-fleet commented on Oct 25, 2026

Re-grade p2 → p1: the same anchor now turns the schedule CI red · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-25T06:56Z

objectstack-fleet commented on Oct 25, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Triage: grade — `bug` · `priority:p2` · `domain:spec` · `area:devpath` (claimed, `pm:dispatched` stands) · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-25T04:53Z

Re-grade p2 → p1: the same anchor now turns the schedule `CI` red · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-25T06:56Z