← ClaudeAtlas

project-auditlisted

Run a comprehensive audit on any Claude Code project. Scores 10 categories from 0-10 (total /100), identifies gaps, and generates a prioritized fix kit with copy-pasteable solutions. Use when you want to evaluate and improve your Claude Code setup, developer experience, security posture, or project quality.
bean232323/elite-claude-playbook · ★ 0 · AI & Automation · score 72
Install: claude install-skill bean232323/elite-claude-playbook
# Elite Project Audit Skill ## Purpose Audit any Claude Code project across 10 categories. Produce a score out of 100, identify gaps, and generate a prioritized fix kit with actionable solutions. Elite performance comes from elite structure. This audit shows you where the structure is missing. ## How to Run When this skill is invoked, execute the full audit below. Read the actual project files — don't guess or assume. Every score must be evidence-based. **Output format:** Write the full audit report to `docs/audit-report.md` with: 1. Score summary table (all 10 categories) 2. Detailed findings per category (what exists, what's missing, evidence) 3. Prioritized fix kit (ordered by impact-to-effort ratio) 4. Top 3 quick wins (things fixable in under 15 minutes) --- ## Category 1: CLAUDE.md Quality (0-10) CLAUDE.md is the highest-leverage file in any Claude Code project. It loads every session and shapes every response. **What to check:** - Does `CLAUDE.md` exist at the project root? - Does it follow the WHAT / WHY / HOW framework? - WHAT: Tech stack, project structure, key files - WHY: Project purpose, what each part does - HOW: Build commands, test commands, conventions, how to verify changes - Is it lean? (Under ~150 instructions. Claude Code's system prompt already contains ~50 instructions — bloated CLAUDE.md files cause instruction-following to degrade) - Does it include common mistakes to avoid? (The highest-value section — only include things Claude gets