Quality Skills

The qual-* family applies a multi-lens quality assessment to code. These skills are coordinated in 3× free parallel lanes followed by a strong synthesis pass.

Skills

Skill ID	Description	Model Class
`qual-review`	Holistic code review: style, correctness, test coverage gaps, readability	`free`
`qual-code-analysis`	Static analysis pass: dead code, unused imports, complexity metrics, naming issues	`free`
`qual-performance`	Performance audit: algorithmic complexity, N+1 queries, memory leaks, render blocking	`cheap`
`qual-security`	Security scan: OWASP Top 10, injection vectors, auth flaws, secret exposure	`strong`
`qual-refactoring-priority`	Ranks refactoring candidates by risk, impact, and technical debt score	`cheap`

Parallel Execution Pattern

Code Input
  ├── qual-review          (free)  → style + correctness
  ├── qual-code-analysis   (free)  → static metrics
  └── qual-performance     (cheap) → perf audit
          │
          ├── qual-security        (strong, independent)
          └── qual-refactoring-priority (cheap, aggregates above)

When to Use

Situation	Skill(s)
Pre-merge code review	`qual-review` + `qual-code-analysis`
Performance regression investigation	`qual-performance`
Security audit before release	`qual-security`
Prioritising tech-debt backlog	`qual-refactoring-priority`
Full quality gate	All five

Instructions That Invoke These Skills

review — primary consumer of all five
refactor — uses qual-refactoring-priority to decide where to start
testing — uses qual-code-analysis to find under-tested paths
research — uses qual-code-analysis + qual-performance for benchmarking
evaluate / benchmark — scoring uses qual-review as baseline

Output Format

Each skill produces a finding table with severity (critical / high / medium / low), file path + line, description, and suggested fix stub.