Imported from danzam98/claude-skills-toolkit (skills/ntm-test-coverage/SKILL.md). Install upstream with npx skills add danzam98/claude-skills-toolkit --skill ntm-test-coverage. Copyright stays with the author.

Test Coverage Audit

Test coverage is not optional and cannot be half-assed. Comprehensive, real tests are a mark of professional craftsmanship and deep respect for the codebase, the team, and the users who depend on this software. We take our work seriously. Half-assing is not acceptable.

If we don't have full test coverage, create a comprehensive and granular set of beads for all missing areas with tasks, subtasks, and dependency structure overlaid with detailed comments.

Prefer real implementations over mocks. Tests that mock everything prove nothing.

Use ultrathink.

Step 1: Discover the Test Setup

Check AGENTS.md and the project's package config to identify:

• Unit test framework and runner (Vitest, Jest, pytest, Go test, etc.)
• Integration/E2E test framework (Playwright, Cypress, etc.)
• Any API or backend test tooling (Miniflare, Supertest, etc.)
• Existing test directories and conventions

./robot files test --json   # if ./robot is available

Step 2: Audit Coverage

Work through each category below. Skip a category only if it genuinely does not apply to this project's architecture — and when you skip, document why. Never skip out of laziness or time pressure.

Unit Tests

• All utility and helper functions
• All form validation logic
• All data transformation and parsing logic
• Business logic and domain rules
• Content or data collection transforms (if applicable)
• Search query parsing (if applicable)
• Animation variant or UI state logic (if applicable)
• Edge cases and error paths for all of the above

Component / UI Tests

(Skip only if project has no UI components)

• All UI components render correctly with varied props
• Props and configuration handled properly
• Accessibility attributes present (labels, ARIA roles, focus management)
• Error states displayed correctly
• Empty and loading states handled
• User interactions behave as expected

E2E / Integration Tests

(Skip only if project has no browser-facing interface)

• Full navigation flow
• Form submissions (with and without JavaScript)
• Search functionality (if applicable)
• Critical user journeys end-to-end
• Authentication flows (if any)
• Mobile viewport testing

API / Backend Tests

(Skip only if project has no backend or API layer)

• All API endpoints return correct responses
• Authentication and authorization enforced
• Error responses have correct status codes and response bodies
• Rate limiting behavior (if applicable)
• Queue processing and async operations (if applicable)
• Database operations and edge cases (if applicable)

Step 3: Create Beads for Gaps

For each area with missing coverage, create a bead:

br create "Tests: <specific area>" --type task --priority 2

Include in each bead description:

• Specific functions, components, or flows to test
• Edge cases to cover
• Expected logging format
• Acceptance criteria (what "done" looks like)

When to Use

• Before major releases
• When coverage feels incomplete
• After significant new features
• When a bug is found that a test should have caught — that gap must be closed

Tips

• Prefer real implementations over mocks — test actual behavior, not your assumptions about it
• Include detailed logging for easier debugging when tests fail
• Group test beads under a parent test-coverage epic for organization
• Check AGENTS.md for any project-specific testing conventions or CI requirements