Prep RoomSign inStart free
Code RoomDeploy script debugging approach
EasyPrep Room Coding #4507

Deploy script debugging approach

Vibe & agenticAlgorithms & data structuresEntry–Mid~12 min

You hit a confusing error from an unfamiliar Bash script in your deploy pipeline and have two reasonable moves: ask a chat assistant to explain what the script does and why it's failing, or point an autonomous agent at the repo to just fix it. The script touches production deploy commands. Which AI approach fits, and why is the other one risky here?

Implement
choose_ai_assist_mode(understands_change: bool, blast_radius: str, reversible: bool, verifiable: bool) → str
Examples
in[false,"high",false,false]out"explain"
in[true,"low",true,true]out"autonomous"
in[true,"high",false,true]out"draft_and_review"
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 12 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.