Anthropic took down thousands of GitHub repos trying to yank its leaked source code — a move the company says was an accident
A copyright notice targeting leaked Claude Code source code wiped out thousands of legitimate forks.
Anthropic, the AI company behind Claude, caused a massive collateral takedown on GitHub this week. The incident began when a software engineer discovered that Anthropic had accidentally included the proprietary source code for its popular Claude Code command-line application in a recent public release. As AI enthusiasts shared and forked the leaked code to analyze how Anthropic harnesses its underlying LLM, the company issued a Digital Millennium Copyright Act (DMCA) takedown notice to GitHub.
The notice, however, was executed with a sledgehammer instead of a scalpel. GitHub's records show it affected roughly 8,100 repositories, including countless legitimate forks of Anthropic's own public Claude Code repository. This sparked immediate backlash from developers who found their projects suddenly blocked. Anthropic's head of Claude Code, Boris Cherny, stated the broad takedown was accidental. The company quickly retracted the bulk of the notices, limiting the action to the single primary repository containing the leaked code and 96 of its direct forks, with GitHub restoring access to all others.
This operational misstep arrives at an inopportune time for Anthropic, which is reportedly planning an initial public offering (IPO). The incident highlights potential compliance and execution risks, as publicly traded companies face intense scrutiny over data security and intellectual property management. Leaking proprietary source code and then mishandling its cleanup could attract unwanted regulatory and shareholder attention, complicating the path to a successful public debut.
- Anthropic's DMCA takedown notice accidentally targeted ~8,100 GitHub repositories, including legitimate forks.
- The action was triggered by the accidental public release of Claude Code's proprietary source code in a recent update.
- The company retracted most notices, restoring access to all but one repo and 96 specific forks containing the leaked code.
Why It Matters
Highlights major operational risks for AI companies, especially those eyeing IPOs, where compliance and precise execution are critical.