A volunteer-led initiative scanned 150 Bitcoin repositories using advanced AI models, identifying over a dozen critical vulnerabilities requiring immediate attention.
The team developed an open-source AI platform to audit Bitcoin software, focusing on wallets, cryptographic libraries, and infrastructure.
Experts report detecting security flaws at a rate of one critical exploit per hour per auditor, prompting rapid vulnerability disclosures to affected projects.
The initiative, led by AnchorWatch CEO Rob Hamilton, invested approximately $20,000 in AI services to deploy a “Bitcoin red team” approach. This strategy employs models like Kimi K3, GPT Sol, Claude Fable/Opus, and GLM 5.2 to simulate attacker perspectives and generate technical reports.
Hamilton emphasized that this comprehensive scan targets core components of the Bitcoin ecosystem, including high-risk infrastructure. He noted the expense but credited the value of proactive security in preventing potential exploits.
Pseudonymous developer Calle highlighted the team’s progress, stating they identified critical vulnerabilities in multiple projects within 12 hours. Daily expenditure reached $10,000 as they prioritized high-impact components.
The project intentionally maintains discretion about specific targets or technical details to avoid alerting malicious actors.
This initiative reflects a broader trend in the crypto sector, where AI is increasingly deployed to identify security flaws rapidly. Recent examples include Anthropic’s detection of a Zcash vulnerability and AI-assisted discovery of issues in Coldcard wallets and Bitcoin bridges like Boltz.
