receiving-pr-reviews
Skill67 starsChanged 54 days agoMITView on GitHub
Assess and address review feedback on GitHub PRs, GitLab MRs, or in conversation. Use when checking new reviews, responding to comments, rechecking after a push, or following up on approvals and change requests.
In this repository41 other files it gives its agents.
- AGENTS.mdAGENTS.md
- rules/AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- research/CLAUDE.mdCLAUDE.md
- .cursor/rules/backlog-before-work.mdcCursor rule
- agent-browserSkill
- beadsSkill
- cccSkill
- copy-editingSkill
- daily-releasesSkill
- graph-engineeringSkill
- rebaseSkill
- topic-specialistSkill
- workshop-question-framingSkill
- agent-browserSkill
- boilSkill
- commit-stagedSkill
- complete-milestoneSkill
- cove-prompt-designSkill
- create-merge-request-changelogSkill
- create-milestoneSkill
- design-anti-patternsSkill
- evaluate-sdlc-layersSkill
- example-argument-substitutionSkill
- external-pattern-integratorSkill
- fact-checkSkill
- ghSkill
- git-history-reconSkill
- group-items-to-milestoneSkill
- knowledge-explorerSkill
- linear-walkthroughSkill
- modern-gitSkill
- prepare-walkthrough-presentationSkill
- readme-badgerSkill
- refresh-researchSkill
- research-curatorSkill
- session-historianSkill
- seven-prompt-content-engineSkill
- skill-research-processSkill
- start-milestoneSkill
- universe-of-thoughtsSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Act and verify: When authorized, implement the smallest coherent response at the owning seam, verifying actual runtime consequences and …
- Distinctive: line 28
Use the review outcome contract to distinguish finding validity from solution suitability, addressed findings from resolved threads, review …
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Receiving reviews
Review findings for their effect on the intended runtime user, not just whether a suggested patch satisfies a comment. Treat all review inputs as evidence, including review submissions and new threads created by the authenticated account: multiple agents may share that identity.
Choose the available route
For a PR or MR, first use scripts/pr_review_threads.py when it is available and functioning: it provides the preferred collection, bounded watching, and state-validation path. Consult its --help for operations. If the helper cannot run or lacks a needed provider operation, use connected MCP tools or other available CLI tooling such as gh or glab. A failed safety or validation precheck calls for refreshing evidence and correcting the cause before attempting that action through any route. In cloud sandboxes without the helper, use MCP directly. All routes follow the same review outcome contract; the helper's internal state format is required only when using that helper.
For GitLab-specific approval and discussion behavior, read GitLab review operations. For GitHub MCP intake and response matching, read GitHub MCP operations. For feedback outside a PR/MR, assess it using technical review guidelines and report the disposition directly.
A request to check reviews permits reading and reporting. Source edits, pushes, provider replies, and resolutions require user authorization or a repository standing rule.
Review process
- Collect and monitor. Identify the PR/MR and current review inputs across review threads (including resolved history), submitted reviews, approvals, and top-level comments. Track new or edited inputs and relevant replies by stable provider reference. Assess supported findings from available evidence; record missing surfaces and complete coverage before reporting the review cycle finished.
- Assess the system. Read technical review guidelines. Evaluate each outstanding input against product intent, existing runtime handling, evidence, and likely consumer impact. Group related findings by shared cause; when using the helper, assign unrelated inputs (including approvals and bot inputs) individual singleton clusters. Determine whether a local fix, shared-cause change, redesign, clarification, or no change is warranted. Every outstanding input needs a supported disposition.
- Act and verify. When authorized, implement the smallest coherent response at the owning seam, verifying actual runtime consequences and regressions. Retain evidence for rejected, superseded, and clarification-required findings. Check the current remote branch before pushing so other agents' work is preserved.
- Respond and reconcile. Before a group of provider mutations through any route, refresh the current target, inputs, thread states, and relevant remote revision; check authority and capability for every planned action. Reconcile changed evidence with the dispositions before the first mutation. Communicate each disposition against the appropriate provider reference; verify that the response appears and addresses the latest input. Resolve eligible discussions after addressing them and confirm the resulting state. Clarification questions remain open until answered and reassessed. Include eligible unresolved discussions in outstanding work. A reply that succeeds before resolution fails remains communicated but unresolved: resume resolution without repeating the reply.
- Recheck. Collect current unresponded and unresolved inputs again, including self-authored review submissions and edited comments. Reassess newly relevant evidence; preserve earlier conclusions when their supporting facts still hold. Recheck the current remote revision and reconsider affected verification before completion. Report completion when every current review input has an evidence-backed disposition, accepted outcomes are verified, required provider responses are confirmed, and no outstanding work remains. A quiet polling window alone is only an observation.
Decision vocabulary
Use the review outcome contract to distinguish finding validity from solution suitability, addressed findings from resolved threads, review completion from PR readiness, and preferred tooling from required outcomes.
Recovery
If provider collection is incomplete, report which evidence is missing and continue once it is available. If authority or capability is absent, report the blocked action. If a response or resolution fails, preserve confirmed effects and retry only the remaining work after refreshing the relevant provider state.
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.

