AI Red Team Flags 85+ Critical Bitcoin Bugs Across 390 Projects

What happened: A volunteer "Bitcoin Red Team" led by AnchorWatch CEO Rob Hamilton and developer Calle used AI models to scan nearly 400 open-source Bitcoin projects, identifying 85 critical and 635 high-severity vulnerabilities. The team spent up to $40,000 in AI compute tokens over a 27.5-hour sprint, triggered by the recent Coldcard exploit. Plans are underway to open-source their AI-powered security platform, but so far no specific Bitcoin Core vulnerabilities have been publicly disclosed due to ongoing coordinated disclosure processes.

Why it matters: The initiative demonstrates both the promise and urgency of AI-assisted security auditing in crypto. The sheer volume of findings—nearly 5,000 issues in less than two days—highlights the scale of latent risk in the ecosystem. While coordinated disclosure is prudent, the lack of immediate detail on Bitcoin Core bugs leaves some uncertainty for users and developers. The move to open-source the AI harness could accelerate industry-wide improvements in code security.

Source: Decrypt