agentleFS
lamtuanvulamtuanvu/pacer/.agents/skills/project-review/SKILL.md

project-review

Skill1 starsChanged 21 days agoApache-2.0View on GitHub

Use when reviewing a plan or change in this repository, or before delivering changed code as complete.

17 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 12 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    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
  1. Very specific:

    Read project-governance, then apply the project's review gate in .project/rules/repository.md (REPO-10).

    line 8
  2. Distinctive:

    Use fresh subagents where supported and authorized, or a separate reviewer/session.

    line 26
  3. Distinctive:

    Check REPO-13 (proposed, amendment 0.15.0.

    line 35
  4. Distinctive:

    Do not publish, push, merge or message others merely to run this skill.

    line 58

“Common” is measured: kept by at least 6 in 10 of the 12 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Project review

Read project-governance, then apply the project's review gate in .project/rules/repository.md (REPO-10). Rules and reviewer criteria remain canonical under .project/; this entry point needs no Claude plugin.

Establish the scope

For a PR, use its actual target base. For local work include staged, unstaged and relevant untracked files. Record exclusions. Identify the base and reviewed head; for uncommitted files also record a content fingerprint and file inventory so HEAD alone cannot misidentify the reviewed revision. A clean diff is not an audit.

Independent review

Give a fresh reviewer the task requirements, scope/base, governance context packet and raw changes, without the implementer's preferred verdict. Select applicable reviewer definitions from the context index and cover repository rules even for docs-only changes. Do not hard-code reviewer counts or language filters.

Use fresh subagents where supported and authorized, or a separate reviewer/session. The independent review required by REPO-10 cannot be replaced by implementer self-review. If unavailable, report the gate UNVERIFIED and deliver a progress handoff, not a completion claim. Specialist reviews may run sequentially.

Reviewers inspect and report; they do not modify code or grant exceptions. Each finding names its rule/source (or advisory status), severity, location, evidence, impact and remedy. Separate introduced defects from pre-existing issues.

Check REPO-13 (proposed, amendment 0.15.0; advisory until adopted): the delivery names each delegation or planner-side agent with its executor or model tier and the reason, and states any REPO-11 exception.

Save reports under .work/reviews/. Include scope, base/head and working-tree fingerprint if applicable, reviewer identity/independence, governance version, actual checks and evidence, unavailable/inapplicable checks, findings, accepted exceptions, and PASS/FAIL/UNVERIFIED verdict. Missing evidence is never PASS.

Resolve and verify

For UI-affecting work, follow the canonical UI review readiness procedure before PR handoff, draft transitions or completion claims.

The implementer fixes confirmed blocking findings, reruns affected checks and applicable CT-11 live validation, and obtains independent review of the fixes. A disputed finding needs evidence and reviewer reassessment; do not silently drop it. Review all changes made after the previous report, including merge resolutions. Repeat until the gate passes; if a blocker cannot be resolved within authorized scope, report it and the needed decision without claiming completion.

Nonblocking suggestions may be deferred explicitly. Existing UI report contracts, waivers and human merge authority remain in force. Before delivery, confirm the reports cover the final content and summarize evidence and remaining limitations. Do not publish, push, merge or message others merely to run this skill.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.