tech
April 1, 2026
Anthropic leaked its own source Code
The company marketing itself as the safety-first AI lab just had its second major security blunder.
TL;DR
- Anthropic's Claude Code source material was leaked publicly for the second time in over a year.
- The leak occurred when an internal debugging file was accidentally bundled into a routine update of Claude Code.
- The exposed code includes the tool's full architecture, unreleased features, and internal model performance data.
- Leaked features include the ability for Claude to review past sessions for improvements, a persistent background assistant, and remote control capabilities.
- Competitors gain a detailed roadmap of Anthropic's development, including plans for longer autonomous tasks, deeper memory, and multi-agent collaboration.
- Anthropic stated it was a release packaging issue caused by human error, not a security breach, and is implementing measures to prevent recurrence.
- The incident raises concerns about Anthropic's operational security, as it markets itself as a safety-first AI lab.
Continue reading the original article