claude-audit

claude-audit

AI-readiness audit plugin for Claude Code. Analyzes any project's Claude Code setup and produces a structured report with prioritized improvements.

Install

/plugin install claude-audit

Skills

/audit-project

Run from inside the project you want to audit:

/audit-project

Produces a report covering:

• Project Profile — detected stack, secrets management, existing Claude setup
• What's Already Good — areas where the project follows best practices
• Structure Gaps — missing CLAUDE.md, hooks, rules, etc. with priority ratings
• Recommended Automations — MCP servers, hooks, subagents, skills, plugins
• Priority Actions — ranked by impact and effort

The report is saved to .artifacts/specs/YYYY-MM-DD-ai-readiness-audit.md.

/claude-audit:ask <question>

Ask any question about Claude Code — CLAUDE.md, memory, hooks, skills, agents, plugins, MCP, settings, permissions, or best practices. Answers are backed by official documentation.

/claude-audit:ask how do hooks work?

Architecture

• agents/claude-code-expert.md — Claude Code expertise subagent with persistent knowledge base
• agents/automation-analyst.md — automation gap analyst subagent
• agents/automation-analyst-refs/ — reference lookup tables for automation recommendations
• agent-memory-seed/ — pre-built knowledge seed, updated by CI every 3 hours
• scripts/ — CI build pipeline for knowledge seed (slugify, validation, significance detection)
• tests/ — pytest tests for build scripts