Guides explain the engineering decisions behind reliable Python products. Start with the outcome you need:
- Use Python well with AI Coding: give a coding agent enough context, constrain the change, test behavior, and review side effects.
- Context budgeting and bounded tasks: maximize agent accuracy with strict attention budgeting, clear contracts, and automated verification.
- Build a Python product that can be changed safely: turn a script into a product with contracts, boundaries, observability, and a release path.
- Modern Python typing in practice: use Protocol, TypedDict, generics, and static type checking as machine-enforced contracts.
- Reliable async Python patterns: avoid event loop blocking, leverage TaskGroup, and manage concurrency with Semaphore.
For task-sized instructions, use the playbooks. For code you can run immediately, use the examples.
