3 sessions on google apps script with a coding agent: the agent failed at least once in 2, and 1 finished without a human stepping in.
Takeaway: I'd start the next session by naming the full scope up front — scraping, pricing, registration backend, and payments — since each one kept surfacing new hidden requirements the agent had to dig for.
Takeaway: Next time I'd lead with 'the malus is making everyone's score negative during the group stage, only show earned points until knockouts' since that was the change that actually mattered to users watching the live tournament.
Takeaway: I'd start the next session the same way — running Bugbot on the branch diff before merging — since it caught a real high-severity bug I'd have missed otherwise.