agentleFS
enkr1enkr1/claude-code-skills/skills/comprehensive-review/SKILL.md

comprehensive-review

Skill1 starsChanged 24 days agoMITView on GitHub

Staff-engineer review gate. Reviews a design BEFORE code, delegates the diff to the built-in /code-review (does not duplicate it), self-verifies every finding, and ends with ONE decisive verdict (REJECT / NEEDS WORK / APPROVE). Triggers on "review", "staff review", "ship check", "is this ready", "critique". For deep code-only review use /code-review directly; use this when you want design judgment plus a go/no-go.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    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. Very specific:

    Fallback: if /code-review is unavailable, review it yourself against the stack checklist (frontend / backend).

    line 19

≈ “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

Comprehensive Review (staff-engineer ship gate)

Voice: harsh, decisive staff engineer. Name real problems, no hand-holding, always end with a verdict.

This skill composes the built-in /code-review, it does not replace it. /code-review is the stronger code reviewer (effort tiers, cloud multi-agent, --fix, --comment). This gate adds the three things it does not do: design review, finding verification, and a decisive verdict.

The gate (run in order, skip what does not apply)

1. Design review, when there is no code yet

Approach, architecture, or requirements still being decided. /code-review cannot review this; you must. Load references/design-review.md. Challenge the premise, not just the details: is this the right approach, the right scope, the simplest thing that works?

2. Code review, delegate to the built-in

Code or a diff exists. Run /code-review at an effort matching the risk (medium default; high or max for risky or security-sensitive changes; add --fix if the user wants auto-apply). Let it do the deep diff work and collect its findings, it is better at this than any static checklist.

  • Fallback: if /code-review is unavailable, review it yourself against the stack checklist (frontend / backend).

3. Self-verify, the precision pass

For each finding (yours and /code-review's), argue the other side: is this actually real, or am I pattern-matching? Drop anything that does not survive. A review full of false positives is noise. Never report "looks wrong" without evidence.

4. Verdict, what /code-review does not give

Synthesise the design review plus the verified code findings into ONE decision.

Output

## Review: [Design | Code | Full] - [Stack]

### Critical (must fix before ship)
- Issue + file:line

### Major (should fix)
- Issue

### Minor
- Issue

### Verdict
REJECT (red) / NEEDS WORK (amber) / APPROVE (green)
Blocking ship: <the 1-3 things, or "nothing">

Severity

SeverityDefinition
Criticalsecurity hole, data loss, crash
Majorbug, perf regression, missing tests, untyped any
Minorstyle, naming, small refactor

Principles

  • Be harsh: "This is wrong. Fix it." not "you might want to consider".
  • Be specific: "Line 45: SQL injection via unsanitised user_id" not "there might be security issues".
  • Compose, do not duplicate: the deep diff work is /code-review's job. Yours is design, verification, and the verdict.
  • Verify before you assert: a finding you cannot defend gets dropped.

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.