Output guardrails
Validates what an automated reviewer or fixer produces before it is posted or committed: schema, policy and safety checks.
Sub-techniques
2
| ID | Name | Description |
|---|---|---|
V1058.001 | Schema validation | Reject malformed structured output and retry. |
V1058.002 | Policy rails | Block outputs that leak secrets, change protected files or exceed scope. |
Craft it
- Define the output schema and validate every response.
- List protected paths and actions the bot may not take.
- Log every blocked output for review.
Reverse engineer it
Signs that a codebase already runs this technique, and where to read how it is configured.
- Guardrail config files or validators around model calls.
Open-source packages
3
- Guardrails AI guardrails-ai/guardrails
- NeMo Guardrails NVIDIA-NeMo/Guardrails
- Instructor 567-labs/instructor