code-qualitylisted

# Purpose Score TypeScript or Python source on **structural** quality: function length, nesting depth, cyclomatic complexity, duplication, code smells, and project-specific anti-patterns. Every finding cites `file:line`; every category gets a 0–100 score. Audit only — this skill never edits code. "Does this follow our conventions?" → `code-standards`, which owns `any`-type detection, naming, and imports. "Is this too long or too deeply nested?" → here. # Rules (never violate) 1. Never edit source — findings and scores only; fixes go to `@implementation-agent`. 2. Every finding carries a `file:line` citation and a severity (Error / Warning). 3. Apply one threshold set per file — TypeScript thresholds never judge Python. 4. Never grep for `any` or missing types; that is `code-standards`' scope. 5. Skip generated, migration, and vendored files; blank and comment-only lines never count toward function length. 6. Emit the machine-readable `QUALITY-SCORE:` header as the report's first line. # Resources - Read `resources/audit-procedure.md` before auditing: the threshold table, the 9-step procedure, the project-specific check list, the scoring formula, the self-check, escalation triggers, and failure modes. - Read `resources/report-format.md` when writing the report: the output template plus two worked examples (a TypeScript route handler, a Python device module). # How to use ## What it does Audits TypeScript or Python source for structural problems — functions th