Anthropic Claude Code Source Leaked on GitHub

Bloomberg reported that Anthropic suffered an accidental release of internal source code tied to Claude Code, its AI-powered coding assistant. Once the code appeared on GitHub, Anthropic sent takedown requests to remove it, but those requests extended further than planned, sweeping up additional repositories beyond the original source. The leak also exposed features not yet publicly announced, compounding the original exposure. Claude Code is one of Anthropic's most strategically significant products. AI coding assistants are now a core enterprise software category, with companies standardizing on tools that integrate into developer workflows for code generation, debugging, review, and documentation. Anthropic is competing directly with OpenAI, GitHub Copilot, and a growing field of specialized tools in this space. Source code exposure in that context is more than a technical incident. It raises questions about internal controls, development security practices, and how much of Anthropic's product roadmap competitors now have access to. The overly broad takedown response added a secondary problem. Removing repositories beyond the original leak signals that Anthropic's incident response process was reactive rather than precise, which can erode trust in organizations that depend on Anthropic's reliability. For developers and businesses building on Claude's API, this incident does not directly affect service availability or model quality. The practical risk is about IP exposure and competitive intelligence, not data loss for users. The broader lesson applies to any company shipping AI products at speed. Fast-moving development cycles increase the risk of accidental exposure, and incident response needs to be as fast and precise as the development itself. Understanding how your toolchain handles source code, secrets management, and third-party code exposure is foundational security work for any serious development team.