Bitcoin Red Team Uncovers Hundreds of Vulnerabilities in Ecosystem Audit

A volunteer team called Bitcoin Red Team has conducted an extensive audit of nearly the entire Bitcoin open-source ecosystem using Chinese AI models. The team, led by a pseudonymous developer known as Calle, used tools from Moonshot AI and Z.ai to scan over 390 projects in August.

The audit uncovered numerous security issues, including 85 critical vulnerabilities, 635 high-severity vulnerabilities, and other lesser-rated concerns. According to Calle, the team uses a combination of AI tools and manual review to identify vulnerabilities in software libraries, wallets, and Lightning Network applications.

Calle noted that using U.S.-developed models posed usage restrictions, which prompted the team to switch to Chinese models that can run locally. The Kimi K3 model, for example, is capable of handling large codebases with minimal human intervention.