Project Audit

project-audit

A comprehensive project audit skill for Claude Code. Analyzes 6 key areas of your project — structure, quality, DX, security, documentation, and dependencies — and delivers prioritized findings with actionable improvement proposals.

Overview

project-audit performs a systematic health check of any software project by examining:

Area	What it checks
Structure	Directory layout, naming conventions, architecture patterns
Quality	Linter/formatter config, tests, error handling, large files, code smells
DX	README quality, setup scripts, CI/CD, dev environment config
Security	Hardcoded secrets, .env exposure, dependency vulnerabilities
Documentation	README completeness, CHANGELOG, API docs, LICENSE
Dependencies	Lock files, outdated packages, unused dependencies, license compatibility

No external tool dependencies — runs entirely with Claude Code's built-in tools (Read, Write, Glob, Grep, Bash).

Installation

Add this plugin to your Claude Code configuration:

# In your project or global settings
claude plugin add github:skaji18/claude-plugins/plugins/project-audit

Or add it to .claude/settings.json:

{
  "plugins": [
    "github:skaji18/claude-plugins/plugins/project-audit"
  ]
}

Usage

/project-audit [project-path] [options]

Parameters

Parameter	Default	Options
PROJECT_PATH	Current directory	Any valid path
--depth	standard	quick / standard / deep
--focus	all	structure / quality / dx / security / docs / deps / all
--out	{PROJECT_PATH}/.audit	Any valid path
--lang	ja	ja / en

Depth Levels

Depth	Scope	Files read	Competitive analysis
quick	Directory structure + config files + README	~20 key files	None
standard	quick + source code sampling + dependencies	~50 key files	Brief comparison
deep	standard + exhaustive file analysis + detailed review	No limit	Detailed comparison

Output

Two files are generated in the output directory:

audit-report.yaml

Structured YAML data for machine consumption and integration with other tools:

meta:
  tool: project-audit
  version: "1.0.0"
  project_path: "/path/to/project"
  depth: standard
  timestamp: "2026-02-01T10:00:00"

project_structure:
  overview: "..."
  languages: [...]
  frameworks: [...]

issues:
  - id: ISS-001
    severity: high
    category: security
    title: "Hardcoded API key detected"

improvement_proposals:
  high: [...]
  medium: [...]
  low: [...]

overall_assessment:
  health_score: "B"
  scores:
    structure: "B+"
    quality: "A-"
    ...

audit-summary.md

Human-readable Markdown report suitable for sharing with team members, clients, or management. Includes:

• Executive summary (non-technical language)
• Overall health score (S/A/B/C/D scale)
• Per-area score breakdown
• Issues sorted by severity
• Prioritized improvement proposals with effort estimates
• Recommended action plan (immediate / short-term / mid-term)

Scoring

Each area is scored on an S-D scale:

Score	Meaning
S	Exemplary. Minimal room for improvement
A	Good. Only minor issues
B	Standard. Some improvements recommended
C	Needs improvement. Multiple significant issues
D	Critical. Immediate action required

Modifiers + / - provide finer granularity (e.g., B+, A-).

Overall score uses weighted average: security (x1.5), quality (x1.2), others (x1.0).

Examples

Quick audit of current directory

/project-audit --depth quick

Performs a fast scan of directory structure and config files. Ideal for a first look at an unfamiliar project.

Standard audit with security focus

/project-audit /path/to/project --focus security,quality

Runs a standard-depth audit focusing on security and code quality areas only.

Deep audit for refactoring planning

/project-audit /path/to/project --depth deep --lang en

Exhaustive analysis of the entire project with competitive comparison. Outputs in English. Best for creating a comprehensive refactoring roadmap.

Natural language input

/project-audit I want to check if this project is ready for production

The AI interprets the intent and runs an appropriate audit (likely standard depth with focus on security and quality).

Comparison with impact-analysis

Aspect	impact-analysis	project-audit
Purpose	Trace impact of code changes	Diagnose overall project health
Method	LSP + BFS mechanical tracing	AI-powered comprehensive analysis
External tools	lsprefs, LSP servers	None
Output	evidence.tsv + summary.md	audit-report.yaml + audit-summary.md
Granularity	Function/method level	Project level

License

See the repository root LICENSE for details.