Agent skill
unit-test-loop
This skill should be used when the user asks to "improve test coverage", "add unit tests", "TDD", "test this module", "write tests for", "increase coverage", "/ut command", or discusses unit testing strategies. Covers the unit test loop workflow, React Testing Library best practices, query priorities, and coverage improvement strategies.
Install this agent skill to your Project
npx add-skill https://github.com/SomtoUgeh/somto-dev-toolkit/tree/main/skills/unit-test-loop
SKILL.md
Unit Test Loop - Coverage Improvement
Current branch: !git branch --show-current 2>/dev/null || echo "not in git repo"
The unit test loop uses a 2-phase workflow with Dex task tracking for persistent, cross-session test coverage improvement.
The 2-Phase Approach
| Phase | Name | Purpose |
|---|---|---|
| 1 | Coverage Analysis | Identify gaps, prioritize files |
| 2 | Dex Handoff | Create epic + tasks from analysis |
After Phase 2, use /complete <task-id> for each test task.
Starting the Loop
/ut "Improve coverage for auth module" # Basic
/ut "Add tests" --target 80% # With target
Phase 1: Coverage Analysis
- Run coverage command to see current state
- Identify files with low coverage
- Prioritize 3-7 test tasks for user-facing behavior
Output: <phase_complete phase="1"/>
Phase 2: Dex Handoff
Create Dex epic with target, then tasks for each gap:
# Create epic
dex create "Unit Test Coverage" --description "Target: 80% coverage"
# For each gap
dex create "Test: login validation" --parent <epic-id> --description "
File: src/auth/login.ts
Current: 45%
Test should verify:
- [ ] Valid credentials succeed
- [ ] Invalid credentials show error
"
Output: <phase_complete phase="2"/> or <promise>UT SETUP COMPLETE</promise>
Working on Tasks
Use Dex + /complete workflow:
dex list --pending # See what's ready
dex start <id> # Start working
/complete <id> # Run reviewers and complete
React Testing Library Patterns
"The more your tests resemble the way your software is used, the more confidence they can give you."
Query Priority (Use in Order)
| Priority | Query | Use Case |
|---|---|---|
| 1 | getByRole |
Default choice, use name option |
| 2 | getByLabelText |
Form fields with labels |
| 3 | getByPlaceholderText |
Only if no label available |
| 4 | getByText |
Non-interactive elements |
| 5 | getByTestId |
Last resort only |
Query Types
| Type | When to Use |
|---|---|
getBy/getAllBy |
Element exists (throws if not found) |
queryBy/queryAllBy |
Only for asserting absence |
findBy/findAllBy |
Async elements (returns Promise) |
Best Practices
// Use screen
screen.getByRole('button', { name: /submit/i })
// Use userEvent.setup()
const user = userEvent.setup()
await user.click(button)
// Use jest-dom matchers
expect(button).toBeDisabled() // Not: expect(button.disabled).toBe(true)
// Avoid act() - RTL handles it
await screen.findByText('Loaded') // Not: act(() => ...)
Quality Standards
- ONE test per task - Focused, reviewable commits
- User-facing behavior only - Test what users depend on
- Quality over quantity - One great test beats ten shallow ones
- No coverage gaming - Use
/* v8 ignore */for untestable code
Command Reference
/ut "prompt" # Start analysis
/ut "prompt" --target 80% # With target
/cancel-ut # Cancel loop
/complete <task-id> # Complete task with reviewers
Related
/complete- Run reviewers and mark Dex task completedex list- View pending tasksdex-workflowskill - Full Dex usage patterns
Didn't find tool you were looking for?