Coverage Audit

Go beyond numeric coverage percentages. Run tests with coverage instrumentation, then classify uncovered code by risk to produce a prioritised list of what needs testing most.

Usage

/coverage-audit [--generate] [--threshold <percent>]

Examples:

/coverage-audit                      # Analyse coverage and prioritise gaps
/coverage-audit --generate           # Analyse gaps then generate tests for top priorities
/coverage-audit --threshold 80       # Flag modules below 80% coverage

Options

• --generate (optional): After analysis, auto-generate tests for the highest-risk uncovered code (chains to /generate-tests)
• --threshold (optional): Minimum coverage percentage to flag modules below (default: no threshold, prioritise by risk)

Process

1. Run Coverage

   • Detect test framework and coverage tool
   • Run tests with coverage instrumentation
   • Parse coverage report (lcov, coverage.py, Istanbul, etc.)

2. Identify Uncovered Code

   • Extract all uncovered lines, branches, and functions
   • Map to source files and functions

3. Classify by Risk

   • Critical: Auth logic, payment processing, data mutation, encryption
   • High: API endpoints, database operations, error handlers
   • Medium: Business logic, validation, data transformation
   • Low: Utility functions, formatters, logging

4. Prioritise

   • Rank by: risk category, code complexity, recent change frequency (git log)
   • Produce top-10 "most dangerous untested functions" list

5. Generate (if --generate)

   • Feed top priorities to /generate-tests workflow
   • Generate tests for highest-risk uncovered code first

Output

# Coverage Audit

**Overall**: 72% line coverage, 58% branch coverage

## Top 10 Untested Critical Paths
| # | File | Function | Risk | Lines |
|---|------|----------|------|-------|
| 1 | src/auth/verify.ts | verifyToken() | Critical | 15 uncovered |
| 2 | src/payments/charge.ts | processPayment() | Critical | 23 uncovered |

## Modules Below Threshold
| Module | Coverage | Gap |
|--------|----------|-----|
| src/auth/ | 45% | 35% below threshold |

Notes

• Requires the project to have a test suite with coverage support
• Risk classification uses heuristics based on file paths, function names, and code patterns
• Pairs with /generate-tests --coverage-gaps for automated gap filling