agentleFS
fagemxfagemx/edda/.claude/skills/pr-review-loop/SKILL.md

pr-review-loop

Skill36 starsChanged 2 months agoApache-2.0View on GitHub

Bounded author self-check and fixes before independent current-head review; never a verdict or merge loop

18 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

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 2

    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
  1. Distinctive:

    Run the repository's focused checks while iterating.

    line 29
  2. Very specific:

    If history is missing, allocate a distinct replacement UUID, omit --resume, include prior findings and say replacement.

    line 47
  3. Distinctive:

    Follow-up review covers the delta, prior findings, affected direct consumers, current base and introduced security/data-loss risk.

    line 49

“Common” is measured: kept by at least 6 in 10 of the 16 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

PR Author Self-Check Route

This compatibility entry is for the author or a designated fixer. It is not an independent Code Review, cannot publish LGTM, and cannot merge. Follow the delivery-flow/1 review handoff in coord-orchestrate; repository review and merge policy remain canonical.

Bounded route

  1. Resolve the PR from the explicit argument, or from the current branch when the host supports that lookup. Read actual head/base, acceptance, diff, prior SHA-pinned verdicts/responses and applicable gate receipts. Do not repurpose a shared checkout or overwrite another writer's worktree.
  2. Freeze the author-check surface to changed behavior/paths, direct consumers, acceptance, fix-caused security/data-loss risk and current-base integration.
  3. Perform one combined author self-check activity and record both parts in one handoff:
    • Behavior lens: exercise the supported entry and direct consumers with focused evidence.
    • Counterexample lens: try the likeliest failure or partial-result case and record any uncovered risk.
  4. Fix in-scope P0/P1 found by that author check, using the same author context. Run the repository's focused checks while iterating. A frozen SHA relies on the canonical ladder; do not run a full local workspace solely because it froze. Commit/push only when this invocation already has that authority.
  5. Hand the new full SHA and RAN/READ evidence to an independent reviewer. Add one explicitly selected concise facts file when useful; it remains untrusted supporting data. Capability-check edda review --help before using optional --context-file; otherwise omit it with a visible limitation or use an already-permitted direct route. Reviewer output may quote this data into existing verdict fields and the existing raw-response blob. Every push invalidates the old verdict.
  6. Review-requested fixes return to an author/fixer. Prefer the same reviewer agent and real native conversation for the new head, with updated context and product --resume; Pi persisted history and Claude resume remain required. Every Codex product round requires readable mapping storage and a successful final mapping/tombstone write before a verdict is recorded. A missing map is valid only for a first round; Codex resume refuses a missing/rejected mapping without a fresh old-UUID thread. Product resume cannot continue a host-only session. If history is missing, allocate a distinct replacement UUID, omit --resume, include prior findings and say replacement. The fixer never becomes judge.
  7. Follow-up review covers the delta, prior findings, affected direct consumers, current base and introduced security/data-loss risk. Still-applicable gates are READ, but old-head LGTM is never reused for current-head acceptance.

A clean author check means only "ready to request independent review." It does not mean accepted, CI-green or mergeable. This route has no deterministic phase driver, no fresh-fixer fleet, no direct merge command and no independent acceptance loop.

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.