# vet

A skill that pushes the agent to inspect its own git diff after each small change, so gaps between your request and the patch show up early.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 1100 installs · ★ 511
- Author: imbue-ai
- License: AGPL-3.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/imbue-ai/vet/tree/HEAD/skills/vet/SKILL.md
- Updated: 2026-09-14
- HTML: https://affpapa.org/en/nevibecode/skills/vet

## What it does

\`vet` turns review into something the agent does on its own. It runs after every coherent chunk of work rather than at the end, and it reads two inputs: the diff against a base commit, and the transcript of the conversation. Lining those up is where it earns its keep. Code can pass your tests and still solve a different problem than the one you described, and this tool is meant to sit beside the test suite rather than in place of it.

The package brings a CLI and a handful of exporter scripts, one per harness. `export_claude_code_session.py` handles Claude Code, with equivalents for Codex, Gemini CLI and OpenCode. Each one dumps a session to a file, and you pass that file back through `--history-loader`. Options that matter: `--base-commit`, `--confidence-threshold` (0.8 by default), `--output-format json` or `github`, and `--quiet` to silence status lines.

Two parts have to be present for any of this to run: the skill files and the `vet` binary from `verify-everything`. Install the binary with `pip install verify-everything`; nothing happens without it on `PATH`. If the direct API call fails because no key is configured, `--agentic` sends the analysis through a local Claude Code, Codex or OpenCode install. Slower, and usually sharper about what it flags.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add imbue-ai/vet --skill vet -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/vet/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/vet/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add imbue-ai/vet --skill vet -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/vet/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add imbue-ai/vet --skill vet -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/vet/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install imbue-ai/vet vet
```

2. **Manually: put the files in place**
- File: `.github/skills/vet/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add imbue-ai/vet --skill vet -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/imbue-ai/vet --path skills/vet
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/vet/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add imbue-ai/vet --skill vet -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/imbue-ai/vet/tree/HEAD/skills/vet/SKILL.md

## When to use it
- catch a mismatch between the request and what the agent built
- review a change right away instead of at the end of the task
- hand findings to CI as `github`-formatted output
- ignore remarks that belong to another agent working in the same repo
- run the review through a local harness when no API key is set

## Example prompts
- Run vet on the current diff with this session's history loaded: `vet "fix login" --history-loader "python3 ~/.claude/skills/vet/scripts/export_claude_code_session.py --session-file <path-to-session.jsonl>"`
- Check the diff with vet before you start the next task
- List only vet findings above 0.9 confidence

## Similar
- [improve-codebase-architecture](https://affpapa.org/en/nevibecode/skills/improve-codebase-architecture.md) — skill · 1163035 installs
- [tdd](https://affpapa.org/en/nevibecode/skills/tdd.md) — skill · 1141062 installs
- [triage](https://affpapa.org/en/nevibecode/skills/triage.md) — skill · 990453 installs
- [codebase-design](https://affpapa.org/en/nevibecode/skills/codebase-design.md) — skill · 850344 installs
- [code-review](https://affpapa.org/en/nevibecode/skills/code-review.md) — skill · 789521 installs
- [implement](https://affpapa.org/en/nevibecode/skills/implement.md) — skill · 774700 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md