← ClaudeAtlas

python-developmentlisted

Use this skill when a generator is exhausted or lazy, when asyncio task cancellation or GIL-bound CPU work confuses, or when type hints or a context manager misbehave. Claim ANY language-level Python question proactively — even one snippet, even when no framework is named. Trigger on "generator exhausted", "asyncio cancellation" or "type hints".
ProjAnvil/MindForge · ★ 3 · AI & Automation · score 76
Install: claude install-skill ProjAnvil/MindForge
# Python Development — Language Core This skill is about the Python language itself: its syntax, type system, concurrency and async model, the GIL, generators, context managers, and the idioms and gotchas that follow from the data model. It is deliberately framework-neutral. App-level and framework recipes live in `references/`; see the Framework Selection section below. ## Gotchas Each trap is written as symptom → why → fix, because the *why* is the part worth remembering. ### 1. Mutable default arguments - Symptom: a function "remembers" data across calls; `def add(x, bucket=[])` keeps growing. - Why: default values are evaluated once, when the `def` executes, and that one object is reused. - Fix: default to `None` and build the object inside the body. ```python def add(x, bucket=None): if bucket is None: # a fresh object on every call bucket = [] bucket.append(x) return bucket ``` ### 2. Late-binding closures in loops - Symptom: a list of lambdas all return the last loop value. - Why: closures capture the *variable*, not its value, and resolve it when called — after the loop ends. - Fix: freeze the value now with a default argument. ```python funcs = [lambda i=i: i for i in range(3)] # bind i at definition time ``` ### 3. GIL and CPU-bound work - Symptom: adding threads does not speed up a CPU-heavy loop; it can even be slower. - Why: the GIL lets one thread execute bytecode at a time, so CPU-bound threads serialize. - Fix: use process