Anthropic inadvertently took down thousands of unrelated GitHub repositories while attempting to remove its leaked source code. The company described the action as an accident, highlighting risks around automated copyright enforcement in the AI sector.
A recent leak of the source code behind Anthropic's Claude Code has uncovered references to previously hidden features and provides insight into the company's future development plans. Notable among these is 'Kairos,' a proposed persistent AI daemon designed for proactive user support based on sustained session memory. The leak offers a rare look into the evolving landscape of advanced large language model tools.
Anthropic has confirmed a second accidental leak of its Claude Code source within a year, raising concerns about its software security practices. The leak, caused by a packaging error, led to rapid online dissemination and subsequent legal actions to mitigate exposure. No customer data was compromised, but repeated incidents could heighten scrutiny of the company’s operational safeguards.
Anthropic’s Claude Code command line interface source code has leaked after an exposed source map file was included in a recent npm package release. The breach provides a comprehensive look at the application’s inner workings, raising concerns about trade secret protection in the competitive AI sector.
Stay informed with curated insights delivered weekly to your inbox.