Key facts
- A Bitcoin security initiative used AI models to scan 150 repositories.
- Over a dozen critical vulnerabilities have been identified across Bitcoin projects.
- The initiative is developing an open-source AI platform for auditing Bitcoin software.
- The effort has cost approximately $20,000 in AI services.
- Vulnerabilities have been found in wallets, cryptographic libraries, and infrastructure.
A volunteer security initiative, operating as a Bitcoin red team, has utilized advanced AI models to scan approximately 150 Bitcoin repositories, uncovering more than a dozen critical vulnerabilities. The group is developing an open-source AI platform to audit Bitcoin software, with developers reporting that the effort is identifying critical exploits across wallets, cryptographic libraries, and infrastructure.
AnchorWatch CEO Rob Hamilton stated that the initiative has spent around $20,000 on AI services, including models from Kimi K3, OpenAI, Anthropic, and Z.ai. He noted that the team has received assistance from OpenAI to manage the 'Cyber Harness' system, which is described as a more expensive but valuable scan for critical parts of the Bitcoin ecosystem.
Pseudonymous Bitcoin developer Calle indicated that the AI-powered review systems are averaging about one critical exploit per person per hour, with several vulnerabilities reported to projects in the past 12 hours. This initiative underscores the increasing use of artificial intelligence in identifying security flaws within the cryptocurrency sector, a trend also seen with previous AI-assisted discoveries in projects like Zcash and vulnerabilities found in the Coldcard wallet and Bitcoin bridge Boltz.