Theorem-backed catalog of AI-agent vulnerability classes. Each entry pairs an invariant spec with vulnerable+patched fixtures, a Semgrep detection rule, and mitigation guidance. Maintained by Viridis Security.
ssrf ai-agents responsible-disclosure security-research ai-security vulnerability-database semgrep prompt-injection llm-security viridis-security
-
Updated
May 13, 2026 - JavaScript