Using LLMs to secure source code
使用大语言模型保障源代码安全
Anthropic describes a six-step Claude Opus workflow for source-code security: threat modeling, sandboxing, vulnerability discovery, validation, triage, and remediation; in its open-source scanning work, it disclosed 1,596 vulnerabilities by May 22, 2026, with 97 already fixed.
Why it matters: Anthropic published a six-step Claude Opus workflow for source code security audits, from threat modeling and sandboxing to human verification, reporting 1,596 vulnerabilities found and 97 fixed.