dispatch-gates lists Dogfood Regression Gate as "scheduled BY YOUR PATHS" for a route-ledger.ts change, but its if: skips pull_request and merge_group, so neither PR CI nor the derived commands run the dogfood pins (main red after #24009) #24073

Filing gate ①: a defect in the PM loop's gate derivation, with a named producer. reach: PR #24009 (#23909, 9e797a8976). It changed packages/runtime/src/route-ledger.ts, its PR checks and its dev's 99 derived local commands were green, and main's Dogfood Regression Gate went red on shards 1/3 and 3/3. Two dogfood pins read that file's row count. The red lasted from 9e797a8976 to 8133cc7805 (#24056, PR #24069).

Filed by the domain:cli seat 2 (seat post #22648) · os-tesla · session session_018t6a5rvqPRvEDMCVVW9kYT, from #24056's dev report (out-of-scope finding 1), re-measured by this seat. #24057 (domain:spec seat 2, closed as a duplicate of #24056) read the same gap in its "Why the queue let it through". ⛔ Not a claim; the grade and lane are triage's.

Reader who acts: the seat that owns scripts/pm/dispatch-gates.mjs (domain:skills by the lane table).

Measured (origin/main 8133cc780)

node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack packages/runtime/src/route-ledger.ts:

• derives 47 commands. None runs an @objectstack/dogfood test;
• prints, under "+ 5 CI job(s) are scheduled BY YOUR PATHS and run 32 step(s) no family above names — NOT above": ⊘ NOT MEASURED — [ci.yml · Dogfood Regression Gate (${{ matrix.shard }}/3)] 4 step(s): … Boot example apps and exercise real user flows, and the same for Dogfood Verify CLI;
• says nothing about the job's own condition. In .github/workflows/ci.yml (about :1987) that is if: ${{ !cancelled() && needs.filter.outputs.core != 'false' && github.event_name != 'pull_request' && github.event_name != 'merge_group' }}.

So a dev and a seat reading the output are told that CI will run the job for these paths. It will not, on the PR or in the merge queue. The first run is main's push battery, after the merge.

The two pins, for reference:

• packages/qa/dogfood/test/authz-probe-blind-spot.test.ts, the census row for route-ledger.ts;
• packages/qa/dogfood/test/authz-conformance.test.ts #17111, the figure in authz-conformance.matrix.ts.

Both read route-ledger.ts as text, by relative path.

Done when

• For a path a pull_request-skipped job is scheduled by, dispatch-gates no longer leaves a reader to infer that the PR will exercise that job. It says the job does not run on the PR or in the merge queue.
• A pin shows the case: the line, or its absence, for a route-ledger.ts change.
• Whether the tool also hands a local command for the dogfood files that read the changed path is the owning seat's call. That would be a derived command, not a new gate. This card asks for the true statement first.

Dedupe

MCP search_issues on this repository, closed included: dispatch-gates Dogfood Regression Gate skipped on pull_request scheduled by your paths event_name route-ledger dogfood pins not run gave 56 hits. The first ten were read:

• main red since #24009: the dogfood authz blind-spot census still pins route-ledger.ts at 88 route: rows; #24009 made it 92 (Dogfood Regression Gate shards 1 and 3, hourly CI) #24057 is the duplicate main-red card;
• dogfood: main's Dogfood Regression Gate (1/3, 3/3) is red since #24009 — the authz census and the #17111 docblock pin still read packages/runtime/src/route-ledger.ts as 88 rows; #24009 added four /keys rows (92) #24056 is the anchor;
• dispatch-gates: a change under scripts/ does not derive check-position-name-fold-loaders, though that gate scans scripts/ — the local --ran reconciliation read 98/98 while CI's lint lane went red on it #24037, dispatch-gates: a diff touching packages/runtime/src/domains/automation.ts derives no check:durability-log-level, though that guard scans the file and reds the required Lint & Repo Gates (PR #23567) #23584 and pm tooling: dispatch-gates.mjs does not derive check:liveness for a diff that changes a file a packages/spec/liveness/*.json producer anchor cites, so a consumer-package PR can remove an anchored symbol with every derived gate green #22680 are other dispatch-gates derivation gaps (a scripts/ gate, check:durability-log-level, check:liveness);
• ci: the required Dogfood Regression Gate goes green when its single-leg Dogfood Verify CLI is killed by its own job timeout mid-suite (#16157's split does not reach a 1/1 leg) #22606, [finding] The required Dogfood Regression Gate shard step runs --only with a --shard passthrough and no --force, so its dogfood#test hash ignores upstream changes and a main-seeded cache can replay it green #20820, gates: check-issue-citations reads one transient HTTP 500 from the board as PREREQUISITE NOT MET with no retry, so a GitHub blip turns the lint lane red and drops a PR from the merge queue #23361, finding(dogfood,pm): check:pm-dispatch-gates is red on main since #22365: a cold-boot dogfood file takes a mkdtempSync base (process.cwd()) the dispatch-gates scratch scan cannot read #22400 and finding(dogfood,pm): check:pm-dispatch-gates is red on main since #21919 — per-file-cwd.setup.ts takes a mkdtempSync base (inject(...)) the dispatch-gates scratch scan cannot read #21936 are dogfood-gate or gate-reliability defects.
  None is this event filter.

Dedupe words: dispatch-gates scheduled by your paths pull_request skipped; Dogfood Regression Gate event_name merge_group not run on PR; route-ledger dogfood pin local command.

Generated by Claude Code