Skip to main content
Back to Vibe Coder Toolkit
Intelligence

Expert Audit

Score your Claude Code setup

npx promptreports-cli audit

Comprehensive audit of CLAUDE.md, .claude/ structure, skills, .env.local, settings.json, plans, memory. 0-100 score.

audit is the expert inspector. Evaluates your entire Claude Code setup against best practices: CLAUDE.md exists and is right-sized, .claude/ has essential docs (LESSONS.md, progress.txt), skills have descriptions and triggers, .env.local has critical vars, settings.local.json is configured, plans and memory are active. Produces a 0-100 score with findings grouped by severity (critical/high/medium/low/info) and expert recommendations.

Prerequisites

  • None — runs on any project

Flags & Options

FlagDescriptionDefault
--jsonJSON output
--pushUpload results to Command Center
--dry-runPreview audit scope

Examples

Score my setup

npx promptreports-cli audit

See your 0-100 score + action items.

Push to dashboard

npx promptreports-cli audit --push

Log score history on the web.

Output

Score box + severity-grouped findings + skills table + expert recommendations + action items.

┌─ Claude Code Expert Audit ──────────────────┐
│ Score: 72/100 (Good)                         │
│ ████████░░░░░                                │
│ CLAUDE.md: 1,847 words                       │
│ Skills: 8 installed, Plans: 3 active         │
│ Env vars: 14 set (5/8 services)              │
└──────────────────────────────────────────────┘

CRITICAL (1)
  ✗ CLAUDE.md missing
    Primary instruction file for every session
    Action: Create CLAUDE.md with project rules

MEDIUM (3)
  ● 3 skills missing descriptions
    Skills without descriptions won't auto-activate
  ● Settings.local.json has deprecated keys
  ● .claude/LESSONS.md not updated in 30+ days

Expert Recommendations
  → Create CLAUDE.md — single most impactful improvement
  → Add 3-5 domain-specific skills for your workflows
  → Update LESSONS.md after every session

What it reads and writes

Reads

  • CLAUDE.md
  • .claude/** (everything)
  • .env.local
  • ~/.claude/projects/
  • ~/.claude/memory/

Writes

Nothing (read-only)

API calls

  • POST /api/swarm/intelligence/audit-results (only with --push)

Free vs Pro usage

Free tier

  • Weekly score check
  • Onboarding checklist (what's missing from a new project)
  • Code review prep (audit before opening PR)

Pro tier

Upgrade
  • Score history chart — are you improving?
  • Team leaderboard (healthy competition)
  • Auto-generated fix PRs
  • Best-practice drift alerts across team

Pro tips

  • Aim for 85+ before your next funding round / review
  • Critical findings break your productivity — fix those first
  • Run monthly to track improvement
Free to try

Install the whole toolkit.

One command unlocks all 23 tools.