Glossary
Guardrails
Also: guardrails
Checks on an agent's input and output, often run on a fast, cheap model: is the agent being asked for something off-limits, does the reply contain what it shouldn't.
Example
In the OpenAI Agents SDK an input guardrail stops a support agent when the user asks it to do their homework.
Where it shows up
- Types of AI agents: where they run and how they're built — Post 4 of 8