Skip to content

Topic

Agent tool safety

Practices for deciding which tool calls an AI agent may make, including confirmation prompts, permission models and gating on declared tool behaviour.

Current clusters