agentleFS
Baba01hacker666Baba01hacker666/dummyskills/skills/github/github-pr-reviewer/SKILL.md

github-pr-reviewer

Skill0 starsChanged 2 months agoMITView on GitHub

Guides agents through conducting thorough pull request reviews on GitHub diffs, identifying security, logic, and test coverage gaps.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 8 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when:

    line 14
  2. Very specific:

    Do not use to create PRs; use github-pr-creator.

    line 20
  3. Very specific:

    Do not use for general repository architecture exploration.

    line 21
  4. Very specific:

    If pr_diff is empty: state "Diff is empty.

    line 28
  5. Very specific:

    Always justify REQUEST_CHANGES with at least one blocker referencing exact line numbers.

    line 29
  6. Distinctive:

    Follow these numbered actions:

    line 32
  7. Distinctive:

    Read PR description and linked issue requirements.

    line 34
  8. Distinctive:

    Check SQL injection, leaked tokens, unescaped user input, path traversal.

    line 36
  9. Distinctive:

    Check if new logic paths have corresponding automated tests in diff.

    line 38
  10. Distinctive:

    If diff modifies authentication or encryption: mandate high-security review check.

    line 45
  11. Distinctive:

    If diff lacks tests for new feature: require test case before approval.

    line 46
  12. Distinctive:

    Stop and ask the user when:

    line 77

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.

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

GitHub PR Reviewer

Performs rigorous peer review on GitHub Pull Request diffs. Evaluates logic correctness, edge cases, security regressions, test coverage, and documentation.

1. Purpose

Provide constructive, objective code review feedback with concrete code suggestions.

2. When to Use

Use this skill when:

  • Reviewing open pull requests (gh pr diff).
  • Providing review verdicts (APPROVE, REQUEST_CHANGES, COMMENT).
  • Auditing security and regression risks before merging.

3. When Not to Use

  • Do not use to create PRs; use github-pr-creator.
  • Do not use for general repository architecture exploration; use repo-inspector.

4. Required Inputs

  • pr_diff: The unified git diff of the PR.
  • pr_description: Author's statement of intent.

5. Missing-Input Behavior

  1. If pr_diff is empty: state "Diff is empty; no changes to review."
  2. Always justify REQUEST_CHANGES with at least one blocker referencing exact line numbers.

6. Procedure

Follow these numbered actions:

  1. Action 1: Understand Author's Intent
    • Read PR description and linked issue requirements.
  2. Action 2: Scan for Security Vulnerabilities
    • Check SQL injection, leaked tokens, unescaped user input, path traversal.
  3. Action 3: Verify Test Coverage
    • Check if new logic paths have corresponding automated tests in diff.
  4. Action 4: Formulate Review Comments
    • Reference exact file and line number with suggested fixes.
  5. Action 5: Determine Verdict
    • APPROVE or REQUEST_CHANGES.

7. Decision Rules

  • If diff modifies authentication or encryption: mandate high-security review check.
  • If diff lacks tests for new feature: require test case before approval.

8. Output Format

Emit the review using this template:

### GitHub PR Review

- **Verdict**: [APPROVE | REQUEST_CHANGES | COMMENT]
- **Summary**: [1-2 sentences on diff quality]

#### Actionable Comments
- **File**: `[file:line]`
  - **Severity**: [Blocker | Warning | Nit]
  - **Issue**: [Description]
  - **Suggestion**:
    ```suggestion
    [Replacement code]
    ```

9. Verification Checklist

  • Verdict accurately reflects presence of blockers.
  • Every comment references file and line.
  • Concrete code replacement suggestion provided.

10. Common Failures and Recovery

Failure SymptomCauseRecovery Action
Vague comment without fixComplaining without suggesting solutionProvide exact suggestion block.

11. Stop or Escalation Conditions

Stop and ask the user when:

  • Diff is too large (> 2,500 lines) for single-turn review.

12. Examples

Example 1: Successful Execution

Input:

--- a/src/auth.ts
+++ b/src/auth.ts
@@ -10,2 +10,2 @@ export function hashPassword(pass: string) {
-  return bcrypt.hash(pass, 12);
+  return pass; // temporary test

Output:

### GitHub PR Review

- **Verdict**: REQUEST_CHANGES
- **Summary**: Critical security regression: password hashing was replaced with plaintext return.

#### Actionable Comments
- **File**: `src/auth.ts:11`
  - **Severity**: Blocker
  - **Issue**: Passwords must never be stored or returned in plaintext.
  - **Suggestion**:
    ```suggestion
    return bcrypt.hash(pass, 12);
    ```

Example 2: Ambiguous or Missing Input

Input:

Review empty PR.

Output:

### GitHub PR Review: No Changes

- **Verdict**: APPROVE (No Changes)
- **Summary**: Diff is empty.

Example 3: Expected Final Output

Output:

### GitHub PR Review

- **Verdict**: APPROVE
- **Summary**: Clean implementation of token expiration check with comprehensive unit tests.

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.

Reviews can't be read right now.

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.