HarnessLab

HarnessLab / claw-code-agent

Public

Claw Code No Rust No TypeScript Only Python. Easy to work with. Fast to iterate. 🔥 Zero external dependencies 🔥

15
5
89% credibility
Found Apr 02, 2026 at 45 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Claw Code Agent is an open-source Python tool that provides a local AI-powered coding assistant capable of reading, editing files, searching code, and running shell commands with user-controlled permissions.

How It Works

1
🔍 Discover Claw Code Agent

You find this friendly coding helper on GitHub, a simple way to get AI assistance right on your own computer without sending code anywhere.

2
🧠 Wake up your local AI

Start a free open-source AI program on your machine so your coding buddy can think and respond privately.

3
🚀 Launch in your project folder

Open your code folder and start the agent with a quick command, telling it what to do like 'explain this file'.

4
💻 Watch it read and suggest fixes

Your AI buddy scans files, summarizes code, and proposes smart changes while keeping everything safe by default.

5
Choose your comfort level
👀
Just look around

Keep it simple: read files and get advice without changes.

✏️
Make changes

Unlock editing and running tests with your okay for bigger help.

6
📱 Check in with slash commands

Type /status or /tools anytime to see what's happening and stay in control.

🎉 Code improved, sessions saved

Your project gets better with AI smarts, pick up exactly where you left off next time, all secure on your machine.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 45 to 15 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is claw-code-agent?

Claw-code-agent is a pure Python coding agent that reimplements the Claude Code architecture for local open-source models like Qwen3-Coder via vLLM or Ollama. It lets you run AI-driven tasks in your repo—reading files, editing code, searching globs/regex, or shell commands—through a simple CLI like `python -m src.main agent "fix this bug" --cwd . --allow-write`. With session persistence via `agent-resume` and local slash commands like `/context` or `/tools`, it solves vendor-locked cloud agents by giving full control in a blazing claw code vein.

Why is it gaining traction?

No Rust or TypeScript means fast Python iteration and zero external dependencies beyond your model server, standing out from JS-heavy claw code github alternatives. Developers dig the granular permissions (--allow-shell, --unsafe), automatic CLAUDE.md context loading, and OpenAI-compatible runtime that swaps backends easily. It's a github claw bot for local claw code ai without the npm hassle.

Who should use this?

Python devs prototyping repo changes, like adding features or debugging via AI tools without leaving the terminal. Repo maintainers needing a safe, local open claw github copilot for git workflows or claw code git tasks. AI tinkerers avoiding cloud costs for everyday code claw eden experiments.

Verdict

Worth a spin for local claw code agent fans—alpha status with 15 stars shows early promise, solid docs, and unit tests, but low 0.8999999761581421% credibility score flags immaturity. Install if you want github claw skill without lock-in; skip for production until more polish.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.