Parallel AI branches same field
Two engineers on your team are each using AI agents to add a new field — `currency` — to the same Python order-processing service, in parallel branches. One agent is adding it to the API/serializers, the other to the database layer and tax calculator. How do you coordinate the two AI-assisted streams so the branches merge cleanly and the field means the same thing in both? What goes wrong if you don't?
contract_drift(spec: str, declarations: list[str]) → list[str]["currency|str|true|USD",["api|currency|str|true|USD","db|currency|str|false|"]]out["db:default","db:required"]["currency|str|true|USD",["api|currency|str|true|USD","tax|currency|str|true|USD"]]out[]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.
Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.