agentleFS
NemesLaszloNemesLaszlo/Agent-Collection/.claude/skills/cross-review/SKILL.md

cross-review

Skill1 starsChanged 16 days agoMITView on GitHub

Run a cross-model code review using both fable and opus in parallel, then merge and deduplicate findings into a combined summary. Use when the user says cross review, dual review, cross-model review, review with both models, or wants a thorough second-opinion review of changes.

27 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use handoff and tdd.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 4 projects that use this

  • GoogleSkill

    tdd

    Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 19 edited versions we read keep these.

  • 9

    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 a cross-model code review using both fable and opus to catch blind spots.

    line 17
  2. Distinctive:

    If no target specified, review all staged and unstaged changes.

    line 21
  3. Distinctive:

    If a file or directory is specified → review that

    line 28
  4. Distinctive:

    If there are staged/unstaged changes → review those

    line 29
  5. Very specific:

    If no changes → review the diff of the last commit (git diff HEAD~1)

    line 30
  6. Distinctive:

    Use the Agent tool to launch TWO review agents in parallel (both in the same message).

    line 36
  7. Distinctive:

    Test coverage gaps and missing evidence for behavior claims

    line 45
  8. Distinctive:

    After both agents complete:

    line 53
  9. Distinctive:

    Combine: all issues from both reviews into one list

    line 55
  10. Distinctive:

    Run both reviews without asking.

    line 90

“Common” is measured: kept by at least 6 in 10 of the 19 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

Context

  • Current branch: !git branch --show-current || true
  • Staged changes: !git diff --cached --stat || true
  • Unstaged changes: !git diff --stat || true
  • Recent commits: !git log --oneline -5 || true

Your Task

Run a cross-model code review using both fable and opus to catch blind spots.

Target: $ARGUMENTS

If no target specified, review all staged and unstaged changes. If no changes exist, review the most recent commit.


Step 1: Identify What to Review

Determine the review scope:

  • If a file or directory is specified → review that
  • If there are staged/unstaged changes → review those
  • If no changes → review the diff of the last commit (git diff HEAD~1)

Collect the full list of files and changes to review.

Step 2: Launch Parallel Reviews

Use the Agent tool to launch TWO review agents in parallel (both in the same message). Both run in the background so the diff never lands in the main context twice.

Agent A (fable): Launch a general-purpose agent with model: fable. Provide it the list of files/changes and instruct it to review for:

  • Clean code principles (naming, function size, duplication, abstraction)
  • Actions vs service-layer separation (/code-structure): domain rules leaking into services, mechanics duplicated across actions
  • Error handling and edge cases
  • Security vulnerabilities (injection, auth, data exposure)
  • Pattern consistency with the existing codebase
  • Performance concerns
  • Test coverage gaps and missing evidence for behavior claims

It must produce a structured list of issues with severity, file path, line number, and description.

Agent B (opus): Launch a general-purpose agent with model: opus. Same files, same review criteria, independent review. This provides a second perspective.

Step 3: Merge Results

After both agents complete:

  1. Combine all issues from both reviews into one list
  2. Deduplicate — match issues that describe the same problem in the same location
  3. Flag consensus — issues found by BOTH models are high confidence
  4. Flag unique — issues found by only ONE model are worth reviewing but lower confidence

Step 4: Produce the Cross-Validation Summary

## Cross-Review Summary

**Scope:** [what was reviewed]
**Models:** fable + opus

### Must Fix (found by both models)
- [Issue] — `file:line` — [description]

### Should Fix (found by one model)
- [Issue] — `file:line` — [found by: fable|opus] — [description]

### Consider
- [Suggestion] — `file:line` — [found by: fable|opus] — [description]

### What Looks Good
- [Positive observations agreed on by both models]

### Stats
- Total issues: X (Y must fix, Z should fix, W consider)
- Consensus issues (both models): X
- Fable-only issues: X
- Opus-only issues: X

### Next Step
Run `/fix-issues` to address all must-fix and should-fix items.

Run both reviews without asking. Two independent reviews are the point of 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.