test-coverage

Test Coverage Command

Analyze test coverage, identify gaps, and generate missing tests.

Instructions

Step 0: Detect Test Runner and Coverage Target

Before running anything, detect the project's test runner and coverage target from config:

# Check for test runner config
cat pyproject.toml 2>/dev/null | grep -A5 'tool.pytest\|tool.coverage\|tool.unittest'
cat setup.cfg 2>/dev/null | grep -A5 'tool:pytest\|coverage'
cat pytest.ini 2>/dev/null
cat package.json 2>/dev/null | grep -E '"test"|"jest"|"mocha"|"vitest"'

• Test runner: use whatever the project configures (pytest, unittest, jest, vitest, etc.). Fall back to pytest only if nothing is configured and pyproject.toml exists.
• Coverage target: read from pyproject.toml ([tool.coverage.report] → fail_under), setup.cfg, or .coveragerc. Fall back to 80% only if no target is configured.

If no test runner is detected, ask the user: "couldn't detect the test runner — what does this project use?"

Step 1: Run Coverage

Use the detected test runner. For pytest:

pytest --cov --cov-report=term-missing -q

Step 2: Analyze Coverage Report

1. Parse the coverage output
2. List files below the coverage target, sorted worst-first
3. For each under-covered file, identify:
   • Untested functions or methods
   • Missing branch coverage (if/else, try/except, early returns)
   • Dead code that inflates the denominator

If pytest/coverage is not available, perform static analysis instead:

1. List all source files and their corresponding test files
2. Flag source files with NO test file
3. For files with tests, list public functions/methods that have no test
4. Estimate coverage percentage based on tested vs untested functions

Testing Priority Classification:

• HIGH: Core business logic, security-sensitive functions, public APIs, complex algorithms, data pipeline stages
• MEDIUM: Utility functions, configuration loading, error handling paths
• LOW: Simple getters/setters, trivial wrappers, logging-only code