Action governance for AI agents: policy, approval boundaries, and audit evidence before tool calls create real side effects.
Just open sourced Enforra, an action governance layer for AI agents.
System prompts are not a security boundary. When an agent can issue refunds, run commands, or send emails, the control point needs to sit before the tool action executes.
Enforra evaluates YAML policy before your application callback runs and returns allow, block, require_approval, or log_only. Local runtime, no network calls, no hosted service.