Repository navigation
Remove agent work feeds and check agent PRs for a Remit line - #1850
Merged
mrbobbytables merged 1 commit intoOct 23, 2026
Merged
Conversation
Three follow-ups to ADR 0008, each closing a path through which the agent fleet kept producing out-of-remit work. Mutation workflow: drop the scheduled full-tree run. Its only output was the list of surviving mutants, which the fleet read as a work queue. The PR lane (changed files only, break threshold 90) is unchanged, and the full-tree report stays available on request through workflow_dispatch. CONTRIBUTING.md and the stryker test prose stop describing a weekly run. AGENTS.md: stop restating the coverage thresholds, the clone-ceiling headroom rule and the "kill the survivors" instruction in the agent entry point. Each of those sentences was a standing invitation to move a metric. The bullets now point at CONTRIBUTING.md and say the gates are frozen regression gates, not targets. AGENTS.md leaves the docs-threshold-parity list because it no longer states the numbers. PR queue hygiene: an agent-authored PR whose description has no `Remit:` line is labelled `needs-human` and told once, in a comment, which effects ADR 0008 recognises. The job does not judge the claim and never removes the label; a maintainer clears it or closes the PR. Which bot logins count as agents is set by AGENT_AUTHORS (default hivecommons-hive[bot]). DRY_RUN is respected and the curl fallback already covers both gh calls. GOVERNANCE.md documents the rule next to the conflict check. The `roadmap` GitHub label was deleted alongside this change. Remit: maintainer-filed work (ADR 0008 follow-up) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: mrbobbytables <mrbobbytables@users.noreply.github.com>
✅ Deploy Preview for endusers ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
Contributor
cc873fb
ℹ️ Sentinel notice — informational; author is trusted (owner), merge is not blockedHive flagged this PR (author @mrbobbytables, head
Hive did not add the |
This was referenced Oct 23, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to ADR 0008 (#1848). Each change closes a path through which the agent fleet kept producing out-of-remit work after the ADR landed.
Remit: maintainer-filed work (ADR 0008 follow-up).
1. Mutation workflow: no scheduled full-tree run
.github/workflows/mutation.ymlloses itsscheduletrigger. The only output of the weekly run was the list of surviving mutants, which the fleet read as a work queue (most of thequalityagent's ~60 recent issues trace back to it). The PR lane is unchanged: it still mutates only the files a PR touches and fails below thebreakscore of 90. The full-tree report remains available on request through "Run workflow".stryker.config.mjs,CONTRIBUTING.mdandtests/stryker-config.test.mjsstop describing a weekly run;breakThresholditself is untouched.2. AGENTS.md: stop restating metrics in the agent entry point
The Build & Test section no longer lists the coverage thresholds, the "lower a ceiling when the gate reports headroom" rule, or "kill the survivors it lists". Each of those sentences was a standing invitation to move a metric. The bullets now point at
CONTRIBUTING.mdand state that the gates are frozen regression gates, not targets.AGENTS.mdleavesDOCS_STATING_THRESHOLDSintests/docs-threshold-parity.test.mjssince it no longer states the numbers (the test's own comment asks for exactly this). The Remit bullet under Guidelines now tells agents to state the effect as aRemit:line in the PR description.The
roadmapGitHub label has been deleted (no remaining workflow or script referenced it).3. PR queue hygiene: Remit check
scripts/pr-queue-hygiene.mjsgains a remit check that runs in the same pass as the conflict check:user.type === 'Bot'and login inAGENT_AUTHORS, defaulthivecommons-hive[bot]) whose body has no line of the formRemit: <effect>is labelledneeds-humanand gets one comment naming the five effects ADR 0008 recognises.MARKER_AUTHORSrule as the conflict markers, so it is posted once per PR and a spoofed copy does not suppress it.DRY_RUNis respected; both writes go throughgh pr edit/gh pr comment, which the existing curl fallback already supports.GOVERNANCE.mddocuments the rule beside the conflict check. New unit tests cover the regex (plain, bold, upper-case, indented; missing, empty, mid-sentence), author matching, nag detection, and everymain()branch (label/nag present or absent, DRY_RUN, human author, compliant agent, customAGENT_AUTHORS, conflicting plus missing remit).Validation
npm run test:unit:coverage:checkpasses (all source-region floors met).npm run check:duplicationat the 245 ceiling (new tests share arunRemithelper so they add no clones).This touches two workflow files and GOVERNANCE.md, so it takes the independent-review path rather than any low-risk class.