Agent skills for test automation

Severity: Optional

Check automatically:

1. Check for Claude/agent skill definitions:

   # Look for skills directory
   ls -la skills/ .claude/skills/ 2>/dev/null
   
   # Look for skill files related to testing
   find . -type f -name "*.skill.*" -o -name "*skill*.md" 2>/dev/null | xargs grep -l -iE "test|lint|crap|coverage" 2>/dev/null
   
   # Check CLAUDE.md for test instructions
   grep -iE "run.*test|test.*command|lint.*command" CLAUDE.md 2>/dev/null
   

2. Verify skills cover key operations:

   • Run tests
   • Check coverage
   • Run linting
   • Check/report CRAP score

Cross-reference with:

• TEST-001 - skills should invoke the same test commands
• TEST-005 - skills for CRAP score checking
• Section 2 (Dependencies) - linting configuration

Pass criteria:

• Agent instructions exist for running tests
• Agent instructions exist for running linting
• (Bonus) Agent instructions for CRAP/complexity checks

Fail criteria:

• No agent-readable instructions for test/lint commands
• CLAUDE.md exists but doesn't mention how to run tests

If no skills found, ask user: "No agent skills found for running tests or linting. If you want AI agents to run these, consider adding instructions to CLAUDE.md or a skills/ directory."

Evidence to capture:

• Skills/instructions found (or none)
• Test command documented for agents
• Lint command documented for agents
• CRAP/complexity command documented (if any)

---