What problem does it solve? Writing typed Python scripts often leads to over-engineered models, silenced checker errors, or unsafe handling of untyped input like JSON and environment values. This Skill provides practical rules for annotating scripts correctly without project-scale type-system research. ## Core Features & Use Cases - Boundary Typing: Treats untrusted input (JSON, plist, environment values) as object and narrows it with isinstance validation instead of Any. - Minimal Data Modeling: Guides the choice between TypedDict, frozen dataclasses, Literal, enums, and plain str plus validation for small scripts. - Honest Error Handling: Discourages cast(), broad unions, and blanket # type: ignore in favor of narrowing, validation, and narrow error-code ignores. - Use Case: When writing a standalone Python script that parses CLI arguments and JSON config, use this Skill to type every function signature, validate input at the edge, and pass the configured checker cleanly. ## Quick Start Ask the AI to add type annotations to your Python script and validate its untyped JSON or CLI inputs following these typing rules.
