Anthropic (ANTHRO) is working to contain fallout after accidentally revealing internal instructions used to direct its AI ...
A file containing part of the source code appears to have been leaked with the recent Claude Code 2.1.88 update.