I ve been working on CTRLRun because I kept coming back to one problem:
AI agents are starting to do real things like refunds, API calls, database changes, and sending messages, but there often isn t much between the agent deciding to do something and the action actually happening.
CTRLRun is an open-source execution-safety layer for AI agents that can change real systems. It adds approvals, policy enforcement, duplicate-action protection, ambiguous outcome handling, and execution receipts before consequential actions run. Built for the messy production cases where a timeout or retry does not necessarily mean the first action failed, and blindly trying again can cause real damage.
