Asked about What is actually being scored

How do you make sure AI-generated code is correct?

An answer that works

Scope the request narrowly enough that the output is reviewable, read the diff rather than the summary, check it against the actual requirement rather than plausibility, and run tests. For anything non-trivial I write the test first so there is something objective to check against.

Reading an answer is not being able to give one. The section that teaches this is What is actually being scored, 3 hours of Working with coding agents.

Learn this properly

Also asked about Working with coding agents

How do you use AI tools when writing code?When do you not use an AI agent?You are given a large task and an AI agent. What do you do first?How do you review generated code?The agent produces working code you do not fully understand. What now?