6 sessions on matplotlib with a coding agent: the agent failed at least once in 3, and 3 finished without a human stepping in.
Takeaway: I'd start with a more targeted prompt pointing straight at the entry-point scripts (paperbanana.py, app.py, generate_paper_figures.py) instead of asking for a full directory sweep, since most of the value came from the agent files anyway.
Takeaway: I'd start with the same broad 'read everything and summarize' prompt next time since this was a pure exploration task with no code changes needed.
Takeaway: Next time I'd check upfront whether the sandbox even allows Bash execution of scripts that hit external APIs, since none of the three approaches (heredoc, /dev/stdin, direct venv binary) got past the permission wall.
Takeaway: This was a straightforward read-through task, so next time I'd just ask for the summary directly since no corrections were needed.
Takeaway: Next time I'd tell it upfront which python environment has matplotlib and paperbanana installed, since it burned a lot of turns discovering that the system python, the project venv, and pip were all broken before it found the paperbanana venv that actually worked.
Takeaway: Next time I'd keep the session shorter and more scoped — splitting 'update the paper text' from 'generate all figures and recompile' into separate sessions so the context window doesn't balloon and kill the final compile step.