verification-before-completion

SkillMonitoring & ops

This skill makes your AI double-check its coding work before telling you a task is finished. Instead of reporting completion right away, your agent verifies that the result is actually correct first. It comes from the superpowers collection of development skills.

Available today. Use it from your connected AI after setup.

Add the skill, then give your AI a coding task as usual. It will check its own work and only tell you the task is done once the result is verified.

Then ask your AI: use the verification-before-completion skill

What your AI can do with it

  • Verify that coding work is correct before declaring the task complete
  • Confirm the result actually works before reporting back to you
  • Catch mistakes in its own code before you have to
  • Avoid claiming a task is finished when it still needs work
  • Apply a verification step to every coding task it handles

What this skill tells your AI

The instructions your AI receives, as published by a5c-ai/babysitter in library/methodologies/cc10x/skills/verification-before-completion/SKILL.md and read by ahel’s review.

Rules

  • Never claim "should work" without evidence
  • Never claim success without exit code 0
  • Never claim a bug is fixed without reproduction failure
  • Always capture and report exit codes
  • Always run the full test suite, not just targeted tests
  • Use timeout guards (timeout 60s) to prevent hanging

When to Use

  • Before completing any BUILD workflow step
  • Before claiming a DEBUG fix is verified
  • Before approving a REVIEW result
  • Before marking any task as done

Agents Used

  • integration-verifier (primary consumer)
  • component-builder (TDD evidence)
  • bug-investigator (fix evidence)

Signals

GitHub stars
2k
Forks
106
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
verification-before-completion
Source
github.com/a5c-ai/babysitter