Comment by dist-epoch

Anthropic surely is well aware. Most likely they asked separate agents multiple times to code review the proof and look for exploits.