100K stars in a day: Claw-code based on leaked Claude Code smashes GitHub record
Claw-code, a clean slate rewrite of leaked Claude Code, is now the fastest growing GitHub repository, reaching 100K stars in one day. Anthropic is trying to contain the genie released from the bottle, with GitHub repositories containing leaked code being taken down due to a DMCA notice.
The full source code of Anthropic’s major CLI tool Claude Code was leaked on Tuesday.
On Wednesday, a repository called claw-code, born as a mirror that later became a complete rewrite of the Claude Code app, first in Python and later in Rust, crossed 100,000 stars on GitHub.
This makes it the fastest-growing repo in history, surpassing even the previous record holder, OpenClaw, which took weeks to reach 100,000 stars.
The “knock off” now has more stars than the original Anthropic’s Claude Code repository (97,600 stars).
The creator of “claw-code,” Sigrid Jin (instructkr), claims that the whole developer community was in a frenzy after the leak. They chose to rewrite the code instead of directly hosting the leaked source, primarily as a precaution against legal action.
“I did what any engineer would do under pressure: I sat down, ported the core features to Python from scratch, and pushed it before the sun came up,” the repository’s readme file reads.
“The result is a clean-room Python rewrite that captures the architectural patterns of Claw Code's agent harness without copying any proprietary source.”
Even though the Python version is “not yet a full runtime-equivalent replacement for the original TypeScript system,” the community behind the “open source” alternative is now working to deliver a Rust version.
Anthropic’s accidental leak of proprietary code was quickly discovered on Tuesday. The copied/forked versions appeared across thousands of repositories on GitHub, and the news was shared by millions of users on social media.
One of the projects, dubbed “free code,” claims to have stripped all telemetry and guardrails and unlocked all experimental features. It has 2,600 stars.
The “unlocked” Claude Code versions allow the use of other AI models, including those from competitors.
“A clean, buildable fork of Anthropic's Claude Code CLI – the terminal-native AI coding agent,” the description reads.
Another repository with 6,100 stars claims to be the original Claude Code executable version.
Some repositories have already been disabled due to DMCA takedowns, which suggest Anthropic is actively trying to contain the damage.
While Anthropic’s DMCA takedown requests target direct mirrors on major platforms, they’re unlikely to reach decentralized code-sharing platforms. It’s also to be seen how the company’s legal actions will fare against code rewrites.
The developers’ community is also actively dissecting the inner workings of the highly popular tool. Machine learning engineer Alex Kim published a detailed write-up of what was found inside.
“Throughout the codebase, there are references to a feature-gated mode called KAIROS. Based on the code paths in main.tsx, it looks like an unreleased autonomous agent mode,” the researcher described yet unreleased autonomous agent mode.
Curious what others think about this story? Contribute your thoughts to the debate below.
It suggests a persistent, always running in the background, Claude assistant.
Claude Code also allegedly includes anti-distillation features that aim to pollute any training data competitors might attempt to extract. The source code even includes a Tamagotchi-style companion system, frustration detection, and other features
Unlock more exclusive Cybernews content on YouTube.
