Night Audit

Your entire codebase.
Audited overnight.

One command before you sleep. 11 parallel agents sweep 14 categories, fix what is safe, and flag what is not. You wake up to a report and a cleaner codebase.

One-time purchase. Instant delivery. Runs 100% local.

14
Categories
11
Agents
25+
Auto-fixes
1
Command
night audit
$ /fullmoon
FULLMOON v3.0.0 Night Audit
MAP 11 repos discovered
SCAN Engineering 42 findings
SCAN Security 7 findings
SCAN Accessibility 19 findings
SCAN Brand & Design 31 findings
FIX 17 auto-fixes applied
VERIFY 0 regressions
Done. Report: fullmoon-report.html
The Math

What 89€ replaces.

2-3 days
A manual pass over 14 categories. One person, one weekend, once.
1,500€+
An agency review. One PDF, no reruns, invoiced per engagement.
89€
FULLMOON. Tonight, and after every sprint. Same depth, zero meetings.

Buy it once, run it forever. It pays for itself the first night. Get FULLMOON - 89€

14 Categories

Everything. Checked.

14 audit categories across 4 disciplines. The same sweep a specialist team would run, minus the team.

01 Code 5 categories

Engineering & Correctness

Dead code, unused imports, type errors, console leaks, loose equality, nesting depth.

Architecture & Structure

File organization, naming conventions, module boundaries, circular dependencies, monorepo hygiene.

Performance

Bundle size, lazy loading, image optimization, render blocking, Core Web Vitals.

Dependencies & Tooling

Outdated packages, unused dependencies, lockfile drift, license compliance, build tool config.

Testing Quality

Coverage gaps, flaky tests, edge cases, mocking strategy, test data management.

02 Design & Experience 3 categories

UX / UI / Design

Empty, error and loading states, spacing scale, color tokens, font loading, SEO meta tags, mobile and touch targets.

Accessibility

WCAG 2.1 AA compliance. Contrast, alt text, ARIA, keyboard nav, focus states, reduced motion.

User Journey & E2E

Landing to conversion flows, onboarding, checkout, error recovery, cross-device testing.

03 Content & Product 2 categories

Content & Copy

Spelling, grammar, brand voice, placeholder text, link health, affiliate disclosures.

Product & Business Logic

Pricing consistency across surfaces, funnel integrity, analytics tracking, email templates, webhooks.

04 Infra & Governance 4 categories

Security

8 CSP checks, exposed secrets, XSS vectors, auth bypass, cookie flags, dependency CVEs.

Reliability & Ops

Error handling, logging, monitoring, deployment config, rollback capability.

Global Rules & Skills

CLAUDE.md accuracy, skill trigger coverage, command docs, memory files, tool permission hygiene.

Enforcement HooksNEW

Hook discovery, script health, coverage analysis, gap detection, effectiveness tracking.

5 Phases

MAP. SCAN. FIX. VERIFY. REPORT.

01

MAP

Discover repos, tech stacks, dependencies, and project boundaries. Build the workspace graph before touching anything.

02

SCAN

11 parallel agents sweep all 14 categories simultaneously. Every file, every rule, every pattern.

03

FIX

Auto-remediate 17+ safe patterns. console.log removal, color fixes, spacing corrections. Never touches architecture or auth.

04

VERIFY

Rebuild, retest, regression check. Every auto-fix is verified. If something breaks, it gets rolled back.

05

REPORT

Generate a living HTML dashboard with full findings, fix history, and journey comparison. Markdown summary included.

Auto-Fix

Safe fixes. Applied automatically.

console.log removal
Before
console.log('debug:', data); fetchUser(id);
After
fetchUser(id);
Color token fix
Before
color: #fff; background: #000;
After
color: #F5F5F7; background: #1f1f21;
Em dash replacement
Before
Fast, reliable — and free.
After
Fast, reliable, and free.
Missing alt text
Before
<img src="hero.webp">
After
<img src="hero.webp" alt="Hero section">
var to const/let upgrade
Before
var count = 0; var name = 'RAXXO';
After
let count = 0; const name = 'RAXXO';

Never auto-fixes architecture, pricing, authentication, or business logic. Those get flagged in the report for manual review.

Living Report

Reports that remember.

Every FULLMOON run generates a living HTML dashboard. It tracks your codebase health over time, compares audit journeys, and shows exactly what improved (and what regressed).

Audit Journey

Compare findings across runs. See your codebase health trend over weeks and months.

Fix Velocity

Track how fast findings get resolved. Auto-fixes vs. manual fixes, broken down by category.

Regression Detection

Spot issues that were fixed but came back. Prevents the same bugs from resurfacing.

Health Trends

Per-category health scores with visual bars. See where your codebase is strong and where it needs work.

FULLMOON Report
Engineering
95%
Security
88%
Accessibility
72%
Brand
100%
Audit Journey
Mar 26
308
findings
Mar 28
58
findings
Mar 29
12
findings
Get FULLMOON

One purchase. Zero subscriptions.

89

One-time payment. No subscription.

Less than one agency hour. Runs every night you want.
  • 14 audit categories
  • 11 parallel agents
  • 25+ auto-fix patterns
  • Living HTML report with journey tracking
  • Unlimited repos per workspace
  • All tech stacks (JS/TS, Rust, Python, Go, static)
  • Regression detection across runs
  • Enforcement hooks (brand, blog, theme)
  • Lifetime updates
  • Personal + commercial license
Get FULLMOON - 89€

Instant delivery via email. Works on macOS, Linux, WSL.

Early access product. By purchasing, you consent to immediate delivery and waive your right of withdrawal for this digital download.

FAQ

Questions?

Everything you need to know before buying.

What do I need to run FULLMOON?

Claude Code (CLI or Desktop App). FULLMOON is a Claude Code skill that runs inside your existing setup. Install it, type /fullmoon, and it handles the rest.

How long does an audit take?

Depends on workspace size. A single repo takes a few minutes. A full workspace with 10+ repos takes 15 to 45 minutes. The 11 parallel agents run simultaneously, so it scales well.

Is it safe? Will it break my code?

Auto-fixes only touch safe, mechanical patterns (console.log removal, color tokens, spelling). Architecture, authentication, pricing, and business logic are never auto-fixed. Those get flagged in the report for your manual review. Every fix is verified with a rebuild and regression check.

What tech stacks does it support?

JavaScript/TypeScript, Rust, Python, Go, and static sites. The audit categories adapt to whatever stack it finds in your workspace. Shopify Liquid, Next.js, SvelteKit, Astro, and more.

Can I run it on just one project?

Yes. Use /fullmoon project my-app to scope the audit to a single repo. You can also run individual categories with /fullmoon category 2 (for Security only, as an example).

Does it phone home or send my code anywhere?

Zero telemetry from FULLMOON itself: no analytics, no tracking, no phone-home. It runs inside your Claude Code session, so your code is shared with the Claude API exactly as in any other Claude Code use, and nowhere else.

Can I customize which categories run?

Yes. Run /fullmoon category N where N is the category number (1 through 11). You can also skip categories by editing the skill config. The living report only includes categories that were actually scanned.