What are AI agent guardrails?

Guardrails are checks that run on an agent's inputs and outputs to catch sensitive data, prompt injection and off-policy content before it causes harm.

Open source + Enterprise

How AgentOven does it

Built-in guardrails cover PII detection, prompt injection, content filtering, topic restriction, regex rules, maximum length and LlamaGuard 3, and you can add your own. The enterprise edition sets default guardrails for a whole workspace.

PII detectionPrompt injectionLlamaGuard 3Workspace defaults
All 14 features →
Agent harnessAgent loopContext engineeringMemory & sessionsTool calling & MCPMulti-agent orchestrationHuman-in-the-loopEvals & test suitesRL environments & worldsModel routingAgent observabilityAGENTS.md & GitOpsBackground agents