Did you vibe-code 5k+ lines of code without thoroughly reviewing all of them? Is your application held together mostly by thoughts, prayers, and a suspicious amount of copium ? Do you run through your entire development page after every agent commit just to check that nothing randomly broke? If yes, I built something for you. Introducing riddlerun: an open-source agentic end-to-end web testing framework that can be run directly from the terminal. All you need is Docker, an API key, and the ability to describe in a coherent sentence how your application is supposed to behave. I’d be especially grateful for feedback, issue reports, and proposals for the future direction of the project.