Self-Test Skill Skill | Agent Skills

Self-Test Skill

Pattern for iterative testing during implementation.

When to Load This Skill

You are implementing code
You need to verify your work before completing
You want to catch issues early

Self-Test Loop

WHILE implementation not complete:
    Write/modify code
        ↓
    Write tests for new code
        ↓
    Run tests
        ↓
    FAIL? → Fix code, retry
        ↓
    Run lint
        ↓
    FAIL? → Fix issues, retry
        ↓
    Run typecheck
        ↓
    FAIL? → Fix types, retry
        ↓
    Continue to next piece

Running Tests

Use project-specific test commands: @.claude/skills/project/run-tests/SKILL.md

Common patterns:

# Run specific test file
npm test -- --testPathPattern={file}
pytest {file} -v

# Run affected tests
npm test -- --changedSince=HEAD

Running Lint/Typecheck

Use project-specific commands: @.claude/skills/project/lint/SKILL.md

Common patterns:

# TypeScript
npx tsc --noEmit
npx eslint {files} --fix

# Python
mypy {files}
ruff check {files} --fix

Before Declaring Pre-Complete

Checklist:

[ ] New code has tests
[ ] All tests pass
[ ] Lint passes
[ ] Typecheck passes
[ ] No console errors/warnings

If ANY fails, status is NOT pre_complete.

Principles

Test as you go - Don't batch at the end
Fix immediately - Don't accumulate failures
Be honest - Report actual status

Agent Skills: Self-Test Skill

Install this agent skill to your local