Cc Review

---
allowed-tools:
  - Read
  - Glob
  - Grep
  - Bash
disable-model-invocation: true
description: Audits Claude Code skills, hooks, agents, and rules in a single pass — structural lint, quality scoring (1-5 across 6 weighted dimensions), and prioritized improvement recommendations. Use when reviewing, auditing, scoring, linting, improving, or validating any customization. Produces one sectioned report with pass/fail checks, per-dimension scores, weighted total, quality tier, and P1-P5 action items.
---

This skill analyzes and reports. It does not modify files — surface findings and let the user decide what to apply.

## Reference Files

- [dimensions.md](references/dimensions.md) — 6 weighted quality dimensions: rubrics, evidence, improvement patterns
- [lint-rules.md](references/lint-rules.md) — Structural validation checks by customization type (PASS/WARN/FAIL)
- [prioritization.md](references/prioritization.md) — P1-P5 impact/effort matrix for recommendations
- [scoring-examples.md](references/scoring-examples.md) — Calibration: real skill assessments with score rationale
- [improvement-examples.md](references/improvement-examples.md) — Before/after fix examples by dimension
- [report-template.md](assets/report-template.md) — Unified 3-phase report format

## Scoping

- Current working directory is the project root
- Look for `<project-root>/.claude/` as the customization directory
- When the project root IS `~/.claude/`, the customization directory is the project root itself
- If a specific file or directory is passed as an argument, review that target directly
- Check `settings.json` for integration validation

## Three Phases

**Phase 1 — Lint**: Structural validation. Each check reports PASS, WARN, or FAIL. Rules in [lint-rules.md](references/lint-rules.md). Lint failures contextualize scores (e.g., missing description → Trigger Coverage = 1).

**Phase 2 — Score**: 6 dimensions table (Effectiveness 28%, Clarity 22%, Best Practices 17%, Documentation 15%, Verification 10%, Trigger Coverage 8%). Quality tiers table (Production Ready 4.5-5.0, Good 3.5-4.4, Needs Work 2.5-3.4, Poor 1.5-2.4, Unusable 1.0-1.4). Rubrics in [dimensions.md](references/dimensions.md).

**Phase 3 — Improve**: Each finding becomes a recommendation with dimension, impact/effort, priority (P1-P5), and specific fix. P1 first. Matrix in [prioritization.md](references/prioritization.md).

## Process

1. Locate target and identify customization type (skill, agent, hook, rule)
2. Run structural lint checks ([lint-rules.md](references/lint-rules.md))
3. Score each dimension ([dimensions.md](references/dimensions.md))
4. Calculate weighted total and determine quality tier
5. Generate prioritized recommendations ([prioritization.md](references/prioritization.md))
6. Produce report ([report-template.md](assets/report-template.md))

## Scoring Principles

- **Be specific** — cite exact text, line numbers, files as evidence
- **Be fair** — consider the customization's intended scope and type
- **Be consistent** — apply the same standards across all customizations
- **Be calibrated** — a 5 is exemplary; see [scoring-examples.md](references/scoring-examples.md) for anchors

## Do not use when

- Reviewing application or project code (not harness files) — use `code-audit`
- Reviewing a specific git diff for bugs — use `diff-review`
- Evaluating config only against a newly-shipped Claude Code release — use `cc-release-review`