Bitcoin Red Team Taps Chinese AI to Hunt Software Flaws

Bitcoin relies on a broad layer of open-source wallets, nodes and infrastructure, but publicly available code is not necessarily reviewed continuously. Bitcoin Red Team, an effort led by developer Calle and AnchorWatch Chief Executive Rob Hamilton, is combining artificial intelligence with human verification to identify vulnerabilities before attackers can exploit them, highlighting AI’s growing role in cryptocurrency security audits.

Calle said on Aug. 5, 2026, that the 16-person team scanned 390 projects and filed 4,962 findings in 27.5 hours, including 85 rated critical and 635 high severity. Early computing costs totaled about $20,000. The team has recently used Moonshot AI’s Kimi K3 for code analysis, while supporting models and researchers help document and validate results. Calle said most critical findings are reproduced through proof-of-concept tests in local regtest environments, though classifications may still change.