Verify

Prove that a change actually works by running the project's checks and reporting the evidence. Use after implementing anything, when the user asks whether a change works, or before declaring a task done.

44 stars 5 forksProductivity & Notes Plugin marketplace

Install

Claude Code plugin
/plugin marketplace add smartwhale8/claude-playbook

The repo is a Claude Code plugin marketplace. Run /plugin afterwards to pick the plugin that includes this skill.

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add smartwhale8/claude-playbook --skill verify
Manual install
git clone --depth 1 https://github.com/smartwhale8/claude-playbook.git /tmp/claude-playbook
mkdir -p ~/.claude/skills
cp -r /tmp/claude-playbook/.claude/skills/verify ~/.claude/skills/verify

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Verify the change

What to verify, if the user said: $ARGUMENTS

Claude stops when the work looks done. "Looks done" is not evidence. This skill produces evidence.

## Run the checks

Run each command the project defines, in this order, and stop at the first failure to fix it before continuing. The commands come from `.claude/CLAUDE.md`. If they are not recorded there, find them in `package.json`, `Makefile`, `pyproject.toml`, or the CI workflow, and then add them to `CLAUDE.md` so the next session does not repeat this search.

1. **Type check.** The compiler catches what the tests miss.
2. **Lint.** Never suppress a rule to get a pass.
3. **Tests.** The full suite, not only the file you touched. A change that passes its own test and breaks two others is not done.
4. **Build.** The build succeeding is a separate fact from the tests passing.

## Behavioural check

Running the suite proves the tests pass. It does not prove the change does what was asked. Add one check that exercises the actual behaviour:

- An API change: call the endpoint and show the request and the response.
- A UI change: run the app, take a screenshot, and compare it with what was asked for.
- A command-line change: run the command and show its output.
- A data change: query the result and show the rows.

## Report

State the outcome first: verified, or not verified.

Then the evidence, as a short table of each command and its result, followed by the output of anything that failed.
Read the full skill on GitHub

Details

Skill name
verify
Path
.claude/skills/verify/SKILL.md
License
MIT
Repo updated
20 September 2026

FAQ

What does the Verify skill do?

Prove that a change actually works by running the project's checks and reporting the evidence. Use after implementing anything, when the user asks whether a change works, or before declaring a task done.

How do I install the Verify skill?

Claude Code plugin: /plugin marketplace add smartwhale8/claude-playbook. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Verify popular?

It lives in smartwhale8/claude-playbook, which has 44 GitHub stars and 5 forks.

Related skills

Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.