Answers,
with the receipts.
Three questions developers ask when a coding agent is doing the work. Each answer is built from recorded engine output or the catalog, not a mockup, and says what it does not establish.
How do you verify an AI coding agent actually did its job?
The agent says done and the tests pass. What that proves, what it does not, and the three checks that tell you whether the work is real.
The tests pass but the code is broken. Do my tests test anything?
A green suite shows the code ran, not that it is right. How to learn whether a specific test would notice a specific failure, and what the receipt looks like.
What does a team invitation feature need to handle?
Seventeen requirements an invite-a-teammate feature owes, from a reviewed catalog, with tiers and criticality, plus the five features an invitation quietly pulls in.