← all skills

qq-audit

The problem: Claude wrote your code, your copy, and your architecture, which is exactly why it can’t audit any of it fairly, it already believes its own intentions were the right call. You need lenses that don’t know what you meant to build, only what actually got built, applied domain by domain.

This is the master command for the 255-framework audit system: 255 expert lenses across 13 quality dimensions. It analyzes your project, picks the right domains in the right order, and runs each serially, fixing issues before moving to the next domain so fixes compound.

Commands


/qq-audit                   — smart routing: picks the right domains for your project
/qq-audit [domains]         — run specific domains (e.g. "ux + copy")
/qq-audit pre-launch        — 8 domains before going live
/qq-audit quick             — UX + Visual + Copy fast check
/qq-audit post-build        — UX + Visual + Copy + Frontend after a feature
/qq-audit deep              — all 13 domains
/qq-audit list              — show all 13 domains with framework counts

The 13 domains

DomainCommandFrameworksExpert lens
UX/qq-audit-ux20Does the brain work well with this?
Product/qq-audit-product20Is this the right product?
Visual UI/qq-audit-visual22Does this look professional?
Content/Copy/qq-audit-copy22Is the writing good?
Frontend/qq-audit-frontend22Is the code well-built?
Performance/qq-audit-performance22Is it fast?
Security/qq-audit-security23Is it safe?
QA/Testing/qq-audit-testing22Is it tested?
Backend/API/qq-audit-backend22Is the server-side solid?
SEO/qq-audit-seo15Can people find it?
DevOps/qq-audit-devops15Is it deployable and resilient?
Data Quality/qq-audit-data15Is the data trustworthy?
Compliance/qq-audit-compliance15Is it legally sound?
Total255

Convergence protocol

Each domain runs up to 3 rounds. Round 1 finds ~60% of issues. Round 2 finds 25% more: fixes from round 1 expose new problems. Round 3 catches the last 10–15%. Move on when a domain hits 9.5/10 or exhausts 3 rounds.


UX: Round 1 = 7.2 → Round 2 = 8.8 → Round 3 = 9.6 ✓ — moving to Visual
Visual: Round 1 = 8.1 → Round 2 = 9.7 ✓ — moving to Copy

Results

Measured on a production Next.js dashboard (170+ components, UX domain):

ScoreFixes
BaselineSUS 57.50
After round 1SUS 77.5100+
After round 2SUS 90.045
After round 3SUS 92.58
Ceiling~93–95Remaining issues are product decisions

Domain sequencing

No point auditing visual design when UX is broken: you’ll redesign those screens anyway. Domains run in dependency order:


Product → UX → Visual → Copy → Frontend → Backend → Performance
       → Security → Testing → SEO → DevOps → Data → Compliance

Part of Operator Skills, a collection of skills for running a real Claude Code setup.