Prep RoomSign inStart free
Code RoomOpen-source PR from AI-generated code
MediumPrep Room Coding #4180

Open-source PR from AI-generated code

Vibe & agenticAlgorithms & data structuresMid–Senior~16 min

You maintain a popular Apache-2.0 open-source CLI. A contributor opens a PR adding a slick argument-parsing module they generated with an AI agent. The code is good, but you have no idea whether it reproduces a chunk of some MIT/BSD/GPL project, and your project's clean license posture is part of its appeal to enterprise users. As maintainer, what's your policy for accepting AI-generated contributions, and what do you ask of this contributor?

Implement
screen_ai_contribution(matched_licenses: list[str], match_scores: list[int], has_attestation: bool, threshold: int) → str
Examples
in[["MIT","BSD-3-Clause"],[92,40],true,80]out"needs_attribution"
in[["GPL-3.0"],[85],true,80]out"reject_copyleft"
What a strong answer looks like

Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.

0:00 of about 16 min

Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.

Which questions mattered is sealed until you submit. Telling you now would just be handing over the edge cases.