GuardrailResult

The result of a guardrail evaluation. Guardrails are safety checks that run at specific phases of the agent loop and can allow, deny, or rewrite content.

Class Diagram

---
title: GuardrailResult
config:
  look: handDrawn
  theme: colorful
  class:
    hideEmptyMembersBox: true
---
classDiagram
    class GuardrailResult {
        +boolean allowed
        +string reason
        +unknown rewrite
    }

Yaml Example

allowed: true
reason: Content is safe

Properties

Name	Type	Description
allowed	boolean	Whether the content passed the guardrail check
reason	string	Explanation of why the content was allowed or denied
rewrite	unknown	Optional rewritten content to replace the original

Factory Methods

The following factory methods are declared via @factory and are generated automatically by the emitter in every language.

• rewrite(rewrite: unknown)
• deny(reason: string)
• allow()