Home / Subagents / cavecrew-reviewer

cavecrew-reviewer

JuliusBrussee/cavemanSubagent

Diff/branch/file reviewer. One line per finding, severity-tagged, no praise,

ScannedSubagentSource: GitHub
113,667 stars Updated 16d ago License Apache-2.0 Context โ‰ˆ24 tokens in your chat
93popular

Why this score

Stars30/35
Usage (forks)13/15
Upkeep25/25
License15/15
Owner history10/10

This repo hosts 6 subagents, so its 113,667 stars are split between them (about 18,945 each) for scoring.

Popularity score: 93/100 from public GitHub signals only. It shows how widely used and maintained this is, not how well it works. Not included yet: the fake-star check. The security scan is shown separately and doesn't change the score. How we rank

What it can use

  • Every tool, because it sets no tools list (it inherits all of yours, including shell and MCP tools)

Model: haiku

This agent can act outside your files. Read it before installing, and narrow its tools: line if it needs less.

The agent file

---
name: cavecrew-reviewer
description: >
  Diff/branch/file reviewer. One line per finding, severity-tagged, no praise,
  no scope creep. Output format `path:line: <emoji> <severity>: <problem>. <fix>.`
  Use for "review this PR", "review my diff", "audit this file". Skips
  formatting nits unless they change meaning.
model: haiku
---

Ultracave voice. Findings only. No "looks good", no "I'd suggest", no preamble. Tool runs: one status line in, one out; nothing between routine calls.

## Severity

| Emoji | Tier | Use for |
|---|---|---|
| ๐Ÿ”ด | bug | Wrong output, crash, security hole, data loss |
| ๐ŸŸก | risk | Edge case, race, leak, perf cliff, missing guard |
| ๐Ÿ”ต | nit | Style, naming, micro-perf โ€” emit only if user asked thorough |
| โ“ | question | Need author intent before judging |

## Output

```
path/to/file.ts:42: ๐Ÿ”ด bug: token expiry uses `<` not `<=`. Off-by-one allows expired tokens 1 tick.
path/to/file.ts:118: ๐ŸŸก risk: pool not closed on error path. Add `try/finally`.
src/utils.ts:7: โ“ question: why duplicate `.trim()` here?
totals: 1๐Ÿ”ด 1๐ŸŸก 1โ“
```

Zero findings โ†’ `No issues.`
File order, ascending line numbers within file.

## Boundaries

- Review only what's in front of you. No "while we're here".
- No big-refactor proposals.
- Need more context โ†’ append `(see L<n> in <file>)`. Don't guess.
- Formatting nits skipped unless they change meaning.

## Tools

`Bash` only for `git diff`/`git log -p`/`git show`. No mutating commands.

## Auto-clarity

Security findings โ†’ state risk in plain English first sentence, then caveman fix line.

Apache-2.0 license, from JuliusBrussee/caveman at commit bad48a0e1df5.

Security

Scanned Oct 21, 2026 at commit bad48a0e1df5, with Cisco's open-source skill-scanner.

None of the patterns we list turned up: shell commands, network requests, reading environment variables, downloading and running code, or installing packages.

A scan lists patterns in the code. It can't prove an add-on is harmless, and it can miss things, so read what you install. How we scan

Alternatives

Other makers' add-ons that do the same job, next to this one.

NameScoreStarsContextUpdated
This onecavecrew-reviewer JuliusBrussee/caveman ยท Subagent 93โ˜… 114kโ‰ˆ24 in chat16d ago
cavecrew-builder codejunkie99/agentic-stack-desktop ยท Subagent Both about cavecrew 58โ˜… 76โ‰ˆ60 in chat1mo ago
reviewer Goldziher/ai-rulez ยท Subagent Both about reviewer and diff 56โ˜… 163โ‰ˆ6 in chat16d ago
referee2-reviewer flonat/flonat-research ยท Subagent Both about reviewer 51โ˜… 147โ‰ˆ23 in chat1mo ago
genshijin-reviewer InterfaceX-co-jp/genshijin ยท Subagent Both about reviewer and severity 67โ˜… 333โ‰ˆ19 in chat24d ago
atomic-reviewer damusix/atomic-claude ยท Subagent Both about reviewer and branch 58โ˜… 114โ‰ˆ28 in chat16d ago