Skip to content
Get packs

Verify Before Claiming Done

Makes the agent run and read real verification commands before saying tests pass, a build works, or work is complete.

Source Download skill
Repository
jnMetaCode/superpowers-zh
Stars
8,444
Licence
MIT
Files
1 file

What it does

This skill is a set of rules that makes an AI agent prove its work before it says the work is finished. Instead of saying tests pass or a bug is fixed based on a guess or an earlier run, the agent must run the relevant check again, read the full output, and report what actually happened. It also lists common excuses and warning phrases (like 'should work now') that the agent should treat as a signal to stop and verify. The skill is written in Chinese and contains only guidance; it does not include any scripts.

Great for

  • Developers who have seen an AI agent claim success on code that later failed
  • Teams that want consistent evidence (test output, build exit codes) before commits or pull requests
  • Anyone delegating coding tasks to sub-agents who needs their reports checked against the actual diff

Skip it if

  • Projects with no test, lint or build commands: the skill cannot verify anything without them, and the agent may just run ad hoc checks
  • Quick throwaway tasks where a full re-run of checks adds time without value
  • Users looking for a tool that performs verification itself: the skill only changes how the agent behaves and runs whatever commands the project already has

Try asking

  • “Before I open this pull request, make sure the tests actually pass and show me the output”
  • “Fix the login bug and confirm the original failing test now passes before you say it's done”
  • “Check the sub-agent's changes in the git diff instead of trusting its success report”
  • “Run the build and report the exit code before we merge”
Step-by-step guide: how to use Verify Before Claiming DoneInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    verification-before-completion.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip verification-before-completion.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip verification-before-completion.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Name the exact verification command you want run (for example, your test or build command) so the agent does not improvise a weaker check.
  • Ask for the full output or failure count in the reply so you can see the evidence yourself.
  • For bug fixes, ask for the original failing test to be rerun, and for regression tests, ask for the red-green check to be shown.

What's inside

  • SKILL.md

Redistributed unmodified from jnMetaCode/superpowers-zh under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.