why use many token when few token do trick. Skill + local proxy for coding agents: shorter answers, smaller tool output, every original recoverable.
About this project
caveman is an open-source agent and automation project on GitHub: why use many token when few token do trick. Skill + local proxy for coding agents: shorter answers, smaller tool output, every original recoverable. It sits in the agent space, where the point is to let a model plan, call tools and carry a task through several steps instead of answering a single question. It is maintained by JuliusBrussee and published as a public repository you can read, fork and run yourself.
On GitHub the repository is tagged ai, anthropic, caveman, claude, claude-code, llm, meme, prompt-engineering, skill and tokens, which is a fair summary of what it covers. The usual pattern is to wire the project into your own stack, hand it tools, files or internal data, and keep the orchestration logic under your control rather than behind a vendor API.
It is written mainly in Go and released under APACHE-2.0, a permissive license — commercial use is fine as long as copyright notices are kept. It is actively maintained: the most recent commit landed on November 18, 2026.
To try it, clone the repository and follow its README; that file is the canonical install guide. There is also a project page at https://docs.caveman.so/caveman for documentation or a hosted demo. Because it runs on your own infrastructure, budget time for environment setup: dependencies, model weights and configuration usually take longer than the first successful run.
Key facts as recorded on GitHub: 121,147 stars; 7,078 forks; primary language Go; license APACHE-2.0; last commit on November 18, 2026; topics ai, anthropic, caveman, claude, claude-code, llm.