Today's Vibe thoughts: Recently, I delegated the coding work to the sandbox agent and the testing work to cc. However, you don't know the specifics of the code written by the agent or even which direction it's going. How do you test it? It's best not to test manually yourself, as it wastes too much time. After covering all the tests, test them in the GitHub build stream, and then let cc monitor gh to trace them.