Bitcoin AI Security Audit Files 4,962 Findings Across 390 Projects
A volunteer Bitcoin Red Team says it has run a large-scale ecosystem security audit using AI agents, filing 4,962 findings across 390 Bitcoin-related projects in about 30 hours. The first status report says 85 findings are critical and 635 high severity, with 14.5% of issues in the serious range and an average of 1.85 serious findings per project. The effort now involves 16 globally distributed contributors, including some automated agents, and 91% of findings came through automated scan intake. Manual review still plays a major role, but the group says letting contributors use different AI-assisted review styles has surfaced different bugs. About 21% of findings have been reproduced with proof-of-concept code. Privacy, coinjoin, swaps, and payment tools showed the highest share of serious issues, while cryptographic libraries and SDKs produced the most total findings. Only 19 projects have been notified so far, and eight findings were marked false positives. The campaign comes amid growing concern that AI can uncover crypto vulnerabilities at machine speed.
