ultraworkers/claw-code
↗ GitHub[Notice] The repo temporarily locked while ownership transfer. in the meantime we maintain on here: https://github.com/ultraworkers/claw-code-parity. The fastest repo in history to surpass 100K stars ⭐. Better Harness Tools that make real things done. Built in Rust using oh-my-codex.
134,794
Stars
101,744
Forks
1,872
Watchers
1,413
Open Issues
Safety Rating B
No hardcoded secrets or traditional malicious code patterns are visible in the static content provided. However, significant concerns exist: (1) the project explicitly originates from a reportedly leaked proprietary codebase, raising serious legal and ethical red flags; (2) the claimed star growth rate (100K+ stars in 2 hours) is physically impossible through organic means and strongly suggests artificial inflation, which is a common tactic used to lend false credibility to dubious repositories; (3) the ownership transfer notice and references to a separate parity repo suggest an unstable and potentially deceptive project provenance. The repository is rated Caution rather than Unsafe because the actual code content visible appears to be thin scaffolding rather than overtly malicious code, and no production secrets or active exfiltration patterns were found. Curators should exercise significant skepticism before publishing.
ℹAI-assisted review, not a professional security audit.
AI Analysis
A cleanroom rewrite/port project claiming to recreate architectural patterns of a leaked proprietary AI coding assistant ('Claw Code', an apparent reference to Claude Code). The repository contains a Python source tree and an in-progress Rust port, focusing on agent harness tooling, MCP orchestration, CLI interfaces, and tool/command frameworks for AI coding workflows. The project was developed using AI orchestration tools and claims to be a clean-room reimplementation rather than a copy of proprietary source.
Use Cases
- Building an open-source harness runtime for AI coding agents
- CLI-based interactive REPL for AI-assisted coding sessions
- MCP orchestration and tool wiring for agent workflows
- Plugin system development for AI coding assistants
- Session state and context management for AI agents
Tags
Security Findings (2)
The repository's README describes the project as originating from a 'leaked' proprietary codebase (referred to as 'Claw Code', apparently a reference to Anthropic's Claude Code). The authors explicitly acknowledge awareness of potential legal issues and describe quickly porting code to avoid association with leaked source. This raises significant intellectual property concerns and the repository may be redistributing or derived from stolen proprietary material.
The star count (134K stars, 101K forks) is anomalous and almost certainly artificial/manipulated — the README itself boasts about being 'the fastest repo in history to surpass 100K stars in just 2 hours,' which is physically implausible and indicative of star-farming or bot activity designed to artificially inflate visibility and credibility.
Project Connections