verify-and-stop Skill for Claude: Check the Work, Then Stop

verify-and-stop makes your coding assistant prove finished work meets the acceptance conditions, then stop.

Installs: 111.8K; on skills.sh
Stars: 116.7K; on GitHub
Files: 1; besides SKILL.md
Updated: Nov 2026; last repo commit

verify-and-stop tells the assistant to turn your acceptance conditions into the smallest set of checks that proves them. It reuses results that are still current for the same repository state and runs focused checks before wider ones. It labels each result as pass, fail, unavailable or blocked.

It loads for validation-only tasks, completion checks and last-mile proof. It does not edit product code unless your request includes fixes, and it adds no polish or unrelated tests once the criteria pass. It reports commands, results and unresolved risk only. It bundles no scripts.

Good for

Developers who want a final pass/fail check before a merge
Reviewers confirming a change meets its acceptance criteria
Anyone who wants verification without extra changes

Install verify-and-stop

Claude Code

Copy this prompt.

Install the verify-and-stop skill in Claude Code. Run this command:

npx skills add https://github.com/juliusbrussee/caveman --skill verify-and-stop

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

From its listing on skills.sh. Source.

What's in the skill

When it loads

Prove existing work meets acceptance conditions without expanding scope. Use for validation-only tasks, completion checks, focused gate runs, and last-mile proof.

Key facts

Author: Julius Brussee (repo owner juliusbrussee)
Repository: juliusbrussee/caveman, folder skills/verify-and-stop
Installs on skills.sh: 111,755 (as of 2026-11-02)
Repo stars: 116,702 (whole repository, as of 2026-11-02)
License: Apache-2.0
Last repo commit: 2026-11-01
Files in the skill folder: 2 files: SKILL.md and agents/openai.yaml
skills.sh audits: Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass

What it includes

Turn acceptance conditions into the fewest checks
Translates acceptance conditions into the smallest sufficient proof set.
Smallest proof set
Skip re-runs when the repo state matches
Reuses still-current results with matching repository state.
Read · Reuse current results
Run narrow checks first, then wider gates
Runs focused checks before wider gates.
Runs code · Focused before wide
Pass, fail, unavailable or blocked, stated exactly
Distinguishes pass, fail, unavailable and blocked exactly.
Four outcomes
No product code changes unless fixes were requested
Does not edit product code unless the verification request includes fixes.
No edits
Stop at proof; report commands, results, open risk
Stops as soon as the acceptance proof is complete and reports commands, results and unresolved risk only.
Stop and report

From the official docs

Permissions

Low risk

It runs your existing checks and edits nothing by default, so the risk is low.

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

It runs your project checks
Verification means running your tests, builds or linters. Check that those commands are safe to run on your machine.

Questions

What does the verify-and-stop skill do?

It makes the assistant verify work with the fewest checks that prove it, and then stop. It does not add polish, cleanup or extra tests after the criteria pass.

How do I install verify-and-stop in Claude Code, the Claude app or Codex?

Use the install box at the top of the page and pick your client. The SKILL.md bundles no scripts and lists no other requirements. The folder also includes an agents/openai.yaml file. The README says your agent picks these work patterns up when a task fits, so you do not need to call it by name.

Does verify-and-stop need an API key or an account?

No. The SKILL.md asks for no key, login or tool. It is plain instructions for the assistant, and it bundles no scripts.

Is verify-and-stop safe to use?

It is a short instruction file and bundles no scripts. What it changes is how the assistant behaves when it edits your project, so review the diff and keep your usual git and test setup. The permissions section below lists what to watch.

How is verify-and-stop different from surgical-patch?

verify-and-stop checks work that already exists and does not edit product code unless you ask for fixes. surgical-patch makes a bug fix and adds regression proof. Use verify-and-stop at the end of a task, or when you only want a pass/fail answer.

Information

updated Nov 2, 2026

Repository: juliusbrussee/caveman
Folder: skills/verify-and-stop

juliusbrussee/caveman skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 4

verify-and-stop SKILL.md · github.com
caveman README · github.com
verify-and-stop openai.yaml · github.com
surgical-patch SKILL.md · github.com