Every CVE we've tracked, indexed, timed, and cross-referenced.

Searchable archive of adversarial-AI vulnerabilities: prompt injection, jailbreaks, tool abuse, supply-chain compromises, and model-theft CVEs across OpenAI, Anthropic, Google, Meta, Microsoft, Vercel, LangChain, xAI, and more.

Filter by severity (Critical / High / Medium / Info), vendor, attack class (Prompt injection, Jailbreak, Supply chain, Tool abuse, Deserialization, RCE), status, and disclosure date.