claw-code-parity  by ultraworkers

Agentic workflow harness engineering

Created 3 days ago

New!

5,669 stars

Top 8.9% on SourcePulse

GitHubView on GitHub
Project Summary

This project addresses the need for a faster, memory-safe harness runtime by providing a clean-room Python rewrite and an in-progress Rust port of "Claude Code." It targets engineers and researchers interested in AI harness engineering, agent systems, and efficient code execution, offering a reimplementation of architectural patterns without direct code copying.

How It Works

The project focuses on a Python rewrite of the core features, aiming to capture the architectural patterns of the original "Claude Code" agent harness. Concurrently, a Rust port is under development to achieve superior speed and memory safety. Development is heavily augmented by oh-my-codex (OmX), a workflow layer built on OpenAI's Codex, which orchestrates parallel code review and persistent execution loops for architect-level verification.

Quick Start & Requirements

  • Primary commands:
    • Render Python porting summary: python3 -m src.main summary
    • Print workspace manifest: python3 -m src.main manifest
    • Run verification: python3 -m unittest discover -s tests -v
  • Prerequisites: Python 3 is required. The use of oh-my-codex implies potential reliance on OpenAI's Codex, though specific API keys or setup are not detailed.
  • Relevant Links: Discord: https://discord.gg/5TUQKqFWd.

Highlighted Details

  • The project is a "clean-room Python rewrite" capturing architectural patterns without copying proprietary source code
Health Check
Last Commit

21 hours ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
0
Star History
6,107 stars in the last 3 days

Explore Similar Projects

Starred by Li Jiang Li Jiang(Coauthor of AutoGen; Engineer at Microsoft), Dan Guido Dan Guido(Cofounder of Trail of Bits), and
7 more.

claw-code by ultraworkers

N/A
167k
Agent harness engineering framework
Created 5 days ago
Updated 23 hours ago
Feedback? Help us improve.