quality-auditor · oakoss — Claude Code skill: what it does, how to install
quality-auditor runs a weighted audit across 12 dimensions, scores each one against cited evidence and hands back a prioritized fix list.
MIT 120 installs ★ 29 updated 2026-08-17
npx skills add oakoss/agent-skills --skill quality-auditor -a claude-codeWhat quality-auditor does
Twelve dimensions, each with a weight, each scored 1 to 10. Code quality, architecture, docs, usability, security and developer experience carry 10% apiece; performance, testing, maintainability and accessibility 8%; CI/CD 5%; innovation 3%. A number with no file, metric or snippet behind it does not survive the audit.
The run moves through four phases: a registry-versus-filesystem check, discovery, evaluation, synthesis. The heavy lifting goes to the Explore, Task and Plan subagents. Bundled references hold pass/warn/fail thresholds, per-dimension rubrics, a report template and an anti-patterns guide.
Fail the phase 0 completeness check and the overall score cannot go past 6/10. Style-only linting and quick PR reads sit outside its remit; a linter covers those. Code written by an AI gets the verification gap protocol, with critic agents and a human on the critical paths.
What it helps with
- score a codebase across 12 weighted dimensions
- catch unverified claims in AI-written code
- back every rating with a file or a metric
- turn audit findings into a prioritized roadmap
- compare two frameworks on one scale
Example prompts for the agent
- Audit the
api/service against all 12 dimensions and cite evidence for every score. - Review what the agent generated in
src/handlers/using the verification gap protocol. - Write the report from
references/audit-report-template.mdand rank the recommendations.
How to install quality-auditor
1Skills CLI (Vercel)
npx skills add oakoss/agent-skills --skill quality-auditor -a claude-code2Manually: put the files in place
~/.claude/skills/quality-auditor/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/quality-auditor/
1Skills CLI (Vercel)
npx skills add oakoss/agent-skills --skill quality-auditor -a codex2Manually: put the files in place
.agents/skills/quality-auditor/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add oakoss/agent-skills --skill quality-auditor -a cursor2Manually: put the files in place
.cursor/skills/quality-auditor/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install oakoss/agent-skills skills/quality-auditor2Manually: put the files in place
.github/skills/quality-auditor/SKILL.md1Skills CLI (Vercel)
npx skills add oakoss/agent-skills --skill quality-auditor -a gemini-cli2Platform CLI
gemini skills install https://github.com/oakoss/agent-skills --path skills/quality-auditorManually: put the files in place
.devin/skills/quality-auditor/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add oakoss/agent-skills --skill quality-auditor -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about quality-auditor
How do I install quality-auditor in Claude Code?
Run npx skills add oakoss/agent-skills --skill quality-auditor -a claude-code. The skill lands in ~/.claude/skills/quality-auditor/SKILL.md.
What does quality-auditor do?
It scores a codebase on 12 weighted dimensions and returns a report with a prioritized improvement roadmap. Every score needs evidence.
Why is the total score capped at 6/10?
Phase 0 failed because the registry and the filesystem disagree. The skill caps the overall result in that case.
Source text
SKILL.md · MIT
--- name: quality-auditor description: 'Code quality gatekeeper and auditor. Enforces strict quality gates, resolves the AI verification gap, and evaluates codebases across 12 critical dimensions with evidence-based scoring. Use when auditing code quality, reviewing AI-generated code, scoring codebases against industry standards, or enforcing pre-commit quality gates. Use for quality audit, code review, codebase evaluation, security assessment, technical debt analysis.' license: MIT metadata: author: oakoss version: '1.1' --- # Quality Auditor ## Overview Evaluates tools, frameworks, systems, and codebases against the highest industry standards across 12 weighted dimensions. Produces evidence-based scores, identifies anti-patterns, and generates prioritized improvement roadmaps. Applies extra scrutiny to AI-generated code through the verification gap protocol, ensuring velocity does not compromise integrity. **When to use:** Auditing code quality, reviewing AI-generated code, scoring codebases against industry benchmarks, enforcing pre-commit quality gates, comparing tools or frameworks, assessing technical debt. **When NOT to use:** Quick code reviews without scoring, style-only linting (use a linter), feature implementation, routine PR reviews that do not require a full audit. ## Quick Reference | Dimension | Weight | What to Evaluate | | -------------------- | ------ | ------------------------------------------------------------------- | | Code Quality | 10% | Structure, patterns, SOLID, duplication, complexity, error handling | | Architecture | 10% | Design, modularity, scalability, coupling/cohesion, API design | | Documentation | 10% | Completeness, clarity, accuracy, examples, troubleshooting | | Usability | 10% | Learning curve, installation ease, error messages, ergonomics | | Performance | 8% | Speed, resource usage, caching, bundle size, Core Web Vitals | | Security | 10% | OWASP Top 10, input validation, auth, secrets, dependencies | | Testing | 8% | Coverage (unit/integration/e2e), quality, automation, organization | | Maintainability | 8% | Technical debt, readability, refactorability, versioning | | Developer Experience | 10% | Setup ease, debugging, tooling, hot reload, IDE integration | | Accessibility | 8% | WCAG compliance, keyboard nav, screen readers, cognitive load | | CI/CD | 5% | Automation, pipelines, deployment, rollback, monitoring | | Innovation | 3% | Novel approaches, forward-thinking design, unique value | ## Audit Phases | Phase | Name | Purpose | | ----- | --------------------- | ----------------------------------------------------------------- | | 0 | Resource Completeness | Verify registry/filesystem parity; audit fails if this fails | | 1 | Discovery | Read docs, examine code, test system, review supporting materials | | 2 | Evaluation | Score each dimension with evidence, strengths, and weaknesses | | 3 | Synthesis | Executive summary, detailed scores, recommendations, risk matrix | ## Scoring Scale | Score | Rating | Meaning | | ----- | ----------- | ------------------------------------------- | | 10 | Exceptional | Industry-leading, sets new standards | | 8-9 | Excellent | Exceeds expectations significantly | | 6-7 | Good | Meets expectations with improvements needed | | 5 | Acceptable | Below average, significant improvements | | 3-4 | Poor | Major gaps and fundamental problems | | 1-2 | Critical | Barely functional or non-functional | ## Common Mistakes | Mistake | Correct Pattern | | ------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | | Giving inflated scores without evidence | Every score must cite specific files, metrics, or code examples as evidence | | Skipping Phase 0 resource completeness check | Always verify registry completeness first; missing resources cap the overall score at 6/10 | | Evaluating only code quality, ignoring dimensions | Score all 12 dimensions with appropriate weights; architecture, security, and DX matter equally | | Accepting superficial "LGTM" reviews | Perform deep semantic audits checking contract integrity, security sanitization, and performance hygiene | | Trusting AI-generated code without verification | Apply the verification gap protocol: critic agents, verifiable goals, human oversight for critical paths | | Proceeding after audit failure without re-audit | Stop, analyze the deviation, remediate, then restart the checklist from step 1 | | Using 10/10 scores without exceptional evidence | Reserve 10/10 for truly industry-leading work; most quality tools score 6-7 | | Surface-level static analysis only | Combine linting with architectural fit checks, risk-based PR categorization, and context-aware validation | ## Delegation - **Discover codebase structure and gather audit evidence**: Use `Explore` agent to survey file organization, dependencies, test coverage, and documentation - **Execute targeted quality checks across dimensions**: Use `Task` agent to run linters, security scanners, performance profilers, and accessibility audits - **Design quality improvement roadmap**: Use `Plan` agent to prioritize quick wins, short-term, and long-term recommendations from audit findings > For stylistic cleanup of AI-generated prose and code (emdash overuse, slop vocabulary, over-commenting, verbose naming), use the `de-slopify` skill. > > If the `usability-tester` skill is available, delegate usability dimension evaluation and user flow validation to it. > Otherwise, recommend: `pnpm dlx skills add oakoss/agent-skills -s usability-tester -a claude-code -y` ## References - [Audit Rubric](references/audit-rubric.md) -- pass/warn/fail thresholds, weighted scoring methodology, automated vs manual checklists, score caps, report format - [Dimension Rubrics](references/dimension-rubrics.md) -- detailed scoring criteria, evidence requirements, and rubric tables for all 12 dimensions - [Audit Report Template](references/audit-report-template.md) -- structured report format, executive summary, recommendations, risk assessment - [Anti-Patterns Guide](references/anti-patterns-guide.md) -- code, architecture, security, testing, and process anti-patterns to identify during audits - [Verification Gap Protocol](references/verification-gap-protocol.md) -- AI code verification methodology, critic agents, rejection protocol, risk-based review strategies
© oakoss · MIT · oakoss/agent-skills
Similar skills
All skills for Code review and refactoring →quality-auditor by other authors comes first.