Awarded for completing Insecure LLM Output Handling and passing its exam, demonstrating the ability to treat generated text as untrusted input at every downstream boundary, apply destination specific injection defences to model output, validate and encode generated content correctly, contain agentic actions against a manipulated output, and run the review and testing practice this risk requires.
Certificates are designed to be checked, not just displayed.
The credential is valid for 12 months from issue. Its public page states the expiry date, and renewing means passing the current exam again — so the credential always reflects the current material.
