SuperAPI AI Directory

quality

Testing & quality

View repository

Updated 10-16 06:23

About this skill

Quality

Run project quality gates and report results.

Inputs

Ask the user for scope:

  • Default (recommended): lint + unit tests (fast feedback)
  • Full: all tests + smoke tests
  • Dry-run: print commands without executing

Workflow

  1. Load commands from project placeholder mappings:
    • {{LINT_COMMAND}}
    • {{RUN_UNIT_TESTS_COMMAND}} (optional)
    • {{RUN_ALL_TESTS_COMMAND}}
    • {{SMOKE_TEST_COMMAND}} (optional)
  1. Execute in order: lint → tests → smoke (if full scope).

  2. Report:

    ## Quality Gate Results
    Scope: <default|full|dry-run>
    
    Results:
    ✅/❌ Lint: passed | failed
    ✅/❌ Tests: X passed, Y failed
    ✅/❌ Smoke: passed | failed | skipped
    
    [If failures: show failing test names and relevant log lines]
    
    Next: Fix failures | Proceed with PR
    

Notes

  • Stop immediately if lint or tests fail.
  • Do not modify test files.
  • For CI/PR workflows, always run at least lint + tests before marking ready.