VE

Selects and executes the most targeted verification steps to confirm code changes, ensuring proven correctness.

Install

mkdir -p .claude/skills/verify-change && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/18458" && unzip -o skill.zip -d .claude/skills/verify-change && rm skill.zip

Installs to .claude/skills/verify-change

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Use when behavior changed and you need to choose the smallest relevant verification first, then report evidence clearly
119 chars✓ has a “when” trigger
Intermediate

Key capabilities

  • Select the most relevant verification path for a change
  • Run repo-local checks first
  • Escalate verification only if needed
  • Report executed evidence clearly
  • Report remaining verification not run
  • Identify the smallest valid proof

How it works

The skill selects and runs the most relevant verification path for a change, prioritizing repo-local checks, and then reports the executed evidence and any remaining unverified aspects.

Inputs & outputs

You give it
A description of a behavior change or bug fix
You get back
A report detailing the selected verification command/flow, its results, and any remaining unverified aspects

When to use verify-change

  • Verifying a bug fix
  • Confirming behavior changes after a refactor
  • Providing proof of fix before PR merge

About this skill

<!-- GENERATED — DO NOT EDIT: generated by ai-kit installer from packages/ai-universal-rules/templates. -->

What I Do

I select and run the most relevant verification path for a change and separate executed evidence from recommendations.

When To Use Me

  • after a bug fix
  • after a behavior change
  • before claiming success
  • when review needs direct verification evidence

Read Alongside

  • docs/ai/capabilities/verify-change/CAPABILITY.md
  • docs/ai/capabilities/verify-change/checklist.md
  • docs/ai/capabilities/verify-change/gotchas.md
  • docs/ai/capabilities/verify-change/examples.md

Project Commands

  • Main verification command: composer test
  • Main build command: none (PHP tooling/docs kit; no compiled build step)
  • Main test command: composer test (composer test:fast for parallel)
  • Preferred narrow-first pattern: start with the narrowest repo-local check and escalate only if needed

Output

  • selected command or flow
  • why it is the smallest valid proof
  • commands run
  • results
  • remaining verification not run

Gotchas

  • do not claim success without running at least one direct check
  • do not report a build success as proof of behavior correctness unless the project says otherwise

When not to use it

  • When behavior has not changed
  • When there is no need for direct verification evidence

Limitations

  • The skill does not claim success without running at least one direct check
  • The skill does not report a build success as proof of behavior correctness
  • The skill prioritizes repo-local checks

How it compares

This skill provides a structured and prioritized approach to verifying code changes, focusing on the smallest valid proof, unlike a general or ad-hoc testing process.

Compared to similar skills

verify-change side by side with the closest alternatives in the catalog.

SkillInstallsUpdatedSafetyDifficulty
verify-change (this skill)023dNo flagsIntermediate
python-testing-patterns772moReviewIntermediate
dependency-upgrade264moReviewIntermediate
test-cases576moNo flagsBeginner

Try saying

Example prompts that trigger this skill in your AI assistant.

Search skills

Search the agent skills registry