Last released Apr 20, 2026
The full-stack safety layer for AI agents. 22 shields including prompt injection, toxicity, hallucination detection, data exfiltration, privilege escalation, and more — in 2 lines of code.
Supported by
