At Prisma AIRS AI Gateway, we believe in helping you make your workflows as modular as possible.
This is useful when:
- You want the same Guardrail checks but want to take different basic actions on them
- Your Guardrail checks definitions are dependent on an upstream task and are updated in code
- You want greater control over how you want to handle Guardrails
With the Raw Guardrails mode, you can achieve all this.
Example of a Raw Guardrail
In this example:
type: Specifies the type of hook, which is guardrail.
name: Gives a name to the guardrail for identification.
checks: Lists the checks that make up the guardrail. Each check includes an id and parameters for the specific conditions to validate.
Configuring Guardrail Actions
In this example,
deny: Is set to TRUE or FALSE
async: Is set to TRUE or FALSE
sequential: Is set to TRUE or FALSE - when TRUE, guardrail checks run one after another instead of in parallel
on_success: Used to pass custom feedback
on_failure: Used to pass custom feedback
You can forward specific client request headers to a guardrail provider by adding forwardHeaders inside the check’s parameters:
forwardHeaders: Optional string[] of client headers to forward to this check’s provider. If omitted, no client headers are forwarded. See Forwarding Headers for restricted headers and configuration details.
Last modified on September 15, 2026