| name | coverage-analysis |
| description | Running coverage tools, interpreting coverage reports, identifying high-risk gaps, and prioritizing which uncovered code to test first.
|
| requires | {"bins":["python"]} |
Coverage analysis for identifying and prioritizing untested code.
When to activate
Use this skill after tests pass, or when asked to analyze coverage gaps
and determine which untested code should be covered next.
Python coverage
- Run:
pytest --cov=<package> --cov-report=term-missing
- The
MISS column shows uncovered line ranges (e.g. 45-52, 78).
- Configure in
.coveragerc or pyproject.toml under [tool.coverage]:
set omit to exclude vendored code, migrations, and test files from
the report.
- Branch coverage: add
--cov-branch to detect branches where only one
path is tested.
- HTML report:
--cov-report=html generates htmlcov/index.html for
visual inspection.
JavaScript coverage
- Jest:
npx jest --coverage -- outputs a table and generates coverage/
directory.
- Vitest:
npx vitest run --coverage -- requires @vitest/coverage-v8 or
@vitest/coverage-istanbul.
- Interpret the table:
% Stmts, % Branch, % Funcs, % Lines, and
Uncovered Line #s.
- Configure in
jest.config.* or vitest.config.* with
coverageThreshold to enforce minimums.
Interpreting results
- Focus on uncovered branches, not just uncovered lines -- a line can be
"covered" while half its branches are not.
- Identify high-risk uncovered code: error handlers, input validators,
authentication checks, and security-sensitive paths.
- Ignore generated code,
__init__.py files that only re-export, and
vendored third-party code.
- Look for patterns: if every error handler is uncovered, that signals a
systematic gap in error-path testing.
Gap prioritization
Rank uncovered code by risk, not by line count:
- Security-sensitive -- authentication, authorization, input
sanitization, cryptographic operations.
- Business logic -- core domain rules, financial calculations, state
transitions.
- Utilities -- parsing, formatting, data transformation helpers.
- Boilerplate -- logging setup, CLI argument wiring, import guards.
Running
Python:
pytest --cov=mypackage --cov-branch --cov-report=term-missing
JavaScript (Jest):
npx jest --coverage --coverageReporters=text
JavaScript (Vitest):
npx vitest run --coverage
MUST
- Report coverage percentage before and after adding tests so progress is
measurable.
- Prioritize gaps by risk -- security-sensitive code first, boilerplate
last.
- Test branches, not just line coverage -- a covered line with an untested
else-branch is still a gap.
MUST NOT
- Do not aim for 100% coverage blindly -- some code (e.g. platform-specific
branches, debug-only paths) may not be worth covering.
- Do not count generated or vendored code in coverage metrics.
- Do not write meaningless tests just to hit coverage targets -- every test
should verify a real behavior.
