caveman-explore (Julius Brussee): a read-only code finder for Claude

caveman-explore is a cheap, read-only helper that finds where relevant code lives and returns only file paths and line ranges.

Installs: 114.5K; on skills.sh
Stars: 116.7K; on GitHub
Files: 2; besides SKILL.md
Updated: Nov 2026; last repo commit

caveman-explore is a read-only repository explorer from the caveman project. A main agent hands it a question about where something lives, and it searches the repo and replies with citations only: one path:START-END line per hit with a short reason. It never edits files, runs commands or proposes a solution.

Use it for cold-start orientation, broad searches across files, or when a direct search failed. Skip it when you already know the file or symbol. Its frontmatter limits it to the Read, Glob and Grep tools and sets the haiku model.

Good for

Developers starting in an unfamiliar repo who need to know where to look
Anyone who wants repo exploration kept out of the main chat context
Teams watching token use on large codebases

Install caveman-explore

Claude Code

Copy this prompt.

Install the caveman-explore skill in Claude Code. Run this command:

npx skills add https://github.com/juliusbrussee/caveman --skill caveman-explore

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

Read-only repository explorer for cold-start orientation, broad cross-file localization, or when a direct search failed. Skip it when the exact file or symbol is already named. Returns path:line citations only; its reads stay out of main context.

Key facts

Author: Julius Brussee
Repo: juliusbrussee/caveman
Skill folder: skills/caveman-explore
skills.sh installs: 114,494
Repo stars: 116,702
License: Apache-2.0
Last repo commit: 2026-11-01
Bundled files: 2 files besides SKILL.md: package.json, tests/skill-file.test.mjs
skills.sh audits: Gen Agent Trust Hub: pass; Socket: pass; Snyk: pass

What it includes

Several Glob, Grep and Read calls in the first turn
Casts a broad net across likely paths, symbols and the most promising files, then follows evidence for one or two more turns if needed.
Files · Parallel search
One line per hit: path/to/file.ext:START-END and a reason
It cites only ranges it actually read. If it finds nothing it replies "no relevant locations found".
Read · Citation-only reply

From the official docs

Permissions

Low risk

The skill reads files in your repo and returns file paths and line ranges. It is limited to read-only tools, so the main risk is that file contents it reads reach your AI provider.

What it can do

Read your data
Create and change things
Delete things
Read-only modeHow to turn it on

Watch out for

It only reads files
Its frontmatter allows Read, Glob and Grep, and SKILL.md says it never edits or runs commands.

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details

Turning on read-only	It is read-only by design: the frontmatter lists only Read, Glob and Grep.
Sign-in and keys	No account or key is named in SKILL.md.
Scopes	Read, Glob and Grep over files in your repository

Questions

What does caveman-explore do?

It searches your repo for where relevant code lives and replies with file paths and line ranges only. Another agent delegates the question and reads just those citations.

How do I install caveman-explore in Claude Code, Claude or Codex?

Use the install box at the top of the page and pick Claude Code, Claude or Codex. Claude Code and Codex read skills from a folder on your machine. In the Claude app you upload the skill folder as a zip.

Can it edit my files?

No. SKILL.md says it never edits files, runs commands or proposes a solution, and its frontmatter allows only the Read, Glob and Grep tools.

Do I need the whole caveman tool?

SKILL.md calls no caveman command. The caveman README lists this skill among those that "Drive the proxy from inside your agent", and offers the whole tool through npx @caveman-ai/cli@latest. The skill is a single prompt you can use on its own.

Does it need an API key?

The skill names no key. It sets the haiku model in its frontmatter.

Is it safe to use?

Its frontmatter allows only Read, Glob and Grep, and SKILL.md says it never edits files or runs commands. It bundles no scripts that run. The lines it reads and returns go to your AI provider like any other tool result.

Information

updated Nov 2, 2026

Repository: juliusbrussee/caveman
Folder: skills/caveman-explore

juliusbrussee/caveman skills.sh

Sources

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

Show all 3

caveman-explore SKILL.md · github.com
caveman README · github.com
caveman-explore on skills.sh · skills.sh