test-coveragelisted

# Purpose Analyze test coverage gaps across the project's repositories (`project.json → repos`) by mapping source files to test files, identifying untested modules and public functions, classifying risk, and suggesting what to test. This skill identifies *what* to test and owns the authoritative coverage-target table; the `testing` skill provides *how* — run this first, then hand off. # Rules (never violate) 1. Read-only: never create or edit test files during gap analysis. 2. Exclude generated files, type declarations, build outputs, and test fixtures from the "untested" count. 3. Every untested module gets a risk classification (High/Medium/Low) with rationale. 4. Label the gap number an estimate — file/function mapping, not tool-measured line coverage. 5. Never run the full suite (`make test-all`) — gap analysis needs no execution. 6. Report stays within 200 lines; over 50 gaps → group by directory, detail the top 10 by risk. # Resources - Read `resources/gap-analysis-procedure.md` when running the analysis: the 6-step procedure (scope, inventory, mapping, gaps, suggested cases, target comparison) with checkpoints, the report template, the authoritative coverage-target table, self-check, mistakes, escalation, and failure modes. # How to use ## What it does Finds the holes in your test suite without touching a single file: it walks a repository or module, matches every source file to its expected test file, and reports what has no coverage — plus which exports insid