monitor-pr
Skill3 starsChanged 46 days agoMITView on GitHub
Monitor one exact GitHub pull request or its managed Stack until ready to merge. Use after initial implementation is complete to assign one persistent worker per PR that marks it ready, monitors current-head CI, conflicts, and configured bot feedback, investigates and fixes valid problems, publishes safely, replies, reacts, resolves threads, and keeps looping until clean or a human decision is required. Never merge.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- microsoft/vscode-team-kittheir own version, first seen Oct 27, 2026
- nexus-a1/claude-skillstheir own version, first seen Oct 17, 2026
- sageox/agent-toolkittheir own version, first seen Oct 8, 2026
- BabylonJS/Babylon.jstheir own version, first seen Oct 8, 2026
- sageox/oxtheir own version, first seen Jun 11, 2026
- doc-detective/agent-toolstheir own version, first seen Apr 10, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- BabylonJS/Babylon.js★ 26ksame name
- sageox/ox★ 67same name
- microsoft/vscode-team-kit★ 65same name
- doc-detective/agent-tools★ 13same name
- sageox/agent-toolkit★ 13same name
- nexus-a1/claude-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository15 other files it gives its agents.
- AGENTS.mdAGENTS.md
- audit-agent-guardrailsSkill
- enforce-codebase-constraintSkill
- protect-gh-repoSkill
- setup-gh-checksSkill
- verify-surfacesSkill
- push-pr-stackSkill
- read-github-issueSkill
- read-github-prSkill
- read-github-resourceSkill
- resolve-pr-threadSkill
- read-imageSkill
- read-text-fileSkill
- read-videoSkill
- transcribe-audioSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Very specific: line 13
When the caller explicitly requests a dry run or read-only evaluation, do not treat invocation as mutation authority.
- Distinctive: line 17
Read the orchestration workflow before acting.
- Distinctive: line 21
Run the bundled snapshot collector to discover the exact PR or managed Stack and record every current head SHA.
- Distinctive: line 25
Keep the coordinator as a control plane only: preserve Stack order and the authoritative head map, schedule bottom-to-top publication …
- Very specific: line 46
Never merge, queue, enable auto-merge, dismiss feedback, bypass requirements, or treat NEUTRAL, skipped, missing, superseded, or stale-head …
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Monitor a pull request to ready
Require one exact https://github.com/OWNER/REPOSITORY/pull/NUMBER URL. Treat invocation as approval to
start review by marking the requested PR, or every open member of its managed Stack, ready. Require the
expected writer login and configured reviewer bot logins from the caller or repository-owned instructions
or configuration; never infer writer authority from the remote owner or guess an actor.
When the caller explicitly requests a dry run or read-only evaluation, do not treat invocation as mutation
authority. Build and report the exact work list, return dry-run, and leave repository and provider state
unchanged.
Read the orchestration workflow before acting.
Coordinate PR owners
- Run the bundled snapshot collector to discover the exact PR or managed Stack and record every current head SHA.
- Assign one persistent sub-agent owner to every scoped PR. Run owners in capacity-limited waves and reactivate the same owner for later events; never split one PR across comment-specific workers.
- Keep the coordinator as a control plane only: preserve Stack order and the authoritative head map, schedule bottom-to-top publication turns, and aggregate worker results. After dispatch, do not inspect technical feedback, edit code, publish branches, or address conversations on a worker's behalf.
- Require each worker to own its complete loop: mark ready, observe only its PR, use
$read-github-pr, reproduce every claim, fix and test valid problems, publish its changes, use$resolve-pr-thread, and resume monitoring the resulting head. - Grant only one exclusive Stack publication turn at a time. The owning worker integrates its own fix,
rebases affected upper branches, tests the resulting Stack, and invokes
$push-pr-stack; the coordinator never performs those steps for it. - Establish readiness only from clean worker reports bound to one final unchanged Stack head map.
Stop condition
Return ready-to-merge only when every scoped PR is ready, linear and conflict-free; every observed
current-head CI and reviewer check and every explicitly configured expected check succeeded; GitHub's
merge state is clean; complete exact-head feedback assessment found no remaining fix, human decision, or
evidence gap; every configured-bot thread is addressed; and no worker remains active. Return the
supported non-ready outcome instead of weakening any condition.
Boundaries
Never merge, queue, enable auto-merge, dismiss feedback, bypass requirements, or treat NEUTRAL, skipped,
missing, superseded, or stale-head checks as success. Serialize Stack rebases and publication, but allow
different workers to investigate and address disjoint threads concurrently. Use one writer per thread.
Pin one shared GitHub CLI configuration and expected actor for the run; verify the actor immediately
before every provider mutation and read the mutation back. An unexpected identity change makes the
operation indeterminate. A lower Stack change invalidates every affected upper worker's assessment;
reactivate that owner at the new head instead of carrying conclusions forward.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

