http-retry-sessionlisted

# HTTP Retry Session Use this skill when a script makes outbound HTTP calls and a single transient failure (a 429, a 503, a dropped connection) should not crash the run - and a slow endpoint should never hang it forever. It wires a `requests.Session` with a `urllib3` `Retry` policy (backoff + status-forcelist + honored `Retry-After`) and a `TimeoutSession` subclass that applies a default connect/read timeout to every request unless overridden. ## When to invoke - User says: "add retries to my requests", "handle 429s", "set a timeout on these calls", "the scraper hangs on a slow URL". - Code in the conversation calls `requests.get`/`post` bare, with no retry policy and no timeout. ## When NOT to invoke - A single fire-and-forget call where a failure should fail loudly and immediately. - You need full async concurrency - reach for `httpx`/`aiohttp` with their own retry/timeout setup instead. ## Concrete example User input: ```text My API client randomly dies on 429s and once hung for ten minutes. Make it resilient. ``` Output: ```python # Copy assets/session.py into your project, then: from session import make_session # BYO base URL via env or config; nothing target-specific is hardcoded. session = make_session(total_retries=4, backoff_factor=0.5, timeout=(3.05, 30)) resp = session.get("https://api.example.com/v1/items", params={"page": 1}) resp.raise_for_status() # retries already exhausted 429/5xx by here data = resp.json() ``` Every call now retries idemp