The problems
- Each problem is a small Python project with planted bugs, bugs are often fixed by changing 1 or 2 lines
- You can solve the problems in any order. Every problem has some easier bugs, so open all of them and try each one before you go deep in a single problem
Scoring
- Each bug you fix gives you points. When you Submit, every bug is checked by its own public and hidden tests
- If your change breaks something that worked before, that Submit scores 0
- Submit as many times as you want. Only your best score on each problem counts
- Your final contest score is your best scores added up
- Submissions after the contest end time do not count
Ties
- A higher total ranks first
- On the same total, fewer Submits ranks first
- If it is still a tie, the one who reached the total first ranks first
Tests
- Run tests runs the tests in
tests/
The agent
An AI agent is in the workspace with you. It has read the rule book, the data files and an outline of the code (e.g. names, signatures and docstrings) but no function bodies. It does not know where the bugs are, cannot run code and does not change your files.
Using any AI tool other than the built-in agent is strictly prohibited. This includes ChatGPT, Copilot, Claude, and any other AI assistant or code completion tool.