Resilient HTTP client retry spec
You need an AI agent to write a resilient HTTP client wrapper (Python) that calls a flaky third-party telecom/SMS API with retries. Write the spec so the retry logic is correct and doesn't make outages worse. What constraints must be explicit, and what does 'add retries with backoff' typically get wrong?
Implement
plan_retry_delays(attempt_outcomes: list[str], is_idempotent: bool, max_attempts: int, total_budget_ms: int) → list[int]Examples
in
[["500","500","ok"],true,5,60000]out[100,200]in
[["timeout","timeout"],false,5,60000]out[]in
[["429:1500","ok"],true,5,60000]out[1500]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 18 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.
Run or narrate your approach, then ask the coach.