Extract Python API surfaces in a compact, token-efficient Markdown format. Parses Python source files using the AST and produces concise summaries containing only classes, functions, and their signatures. Typically reduces token count by 60–90% compared to reading full source files. Supports single-file and whole-project summarisation with options to strip docstrings, exclude private methods, filter to __all__ exports, wrap long signatures, limit recursion depth, and exclude paths by glob pattern.
Extract and summarise Python API surfaces using api-squash. Use when the user asks to summarise a Python project's API, understand a codebase's class or function structure, list public interfaces, inspect module signatures, or before code-generation tasks that need structural awareness — even if they don't say 'api-squash' explicitly. Also use when the user says 'what does this project expose?', 'show me the API surface', or asks to generate code that must conform to an existing Python interface. Do NOT use for non-Python projects or for reading full source code.
Load Python API surfaces into context using api-squash before working on code. Provides a structured workflow for gathering codebase structure — class hierarchies, function signatures, and module layout — so you can reason about a Python project without reading every source file. Covers scope selection (file vs project), flag selection (docstrings, private methods, public-only filtering, signature wrapping, depth), and when to drill deeper.
Load Python API surfaces into context before working on code. Use when you need to understand a Python project's structure before generating code, writing tests, reviewing a PR, planning a refactor, or answering architecture questions — even if the user doesn't ask for a summary explicitly. Also use when starting work in an unfamiliar Python codebase, when the user says 'look at the code first', or when you need to know what classes and functions exist before making changes. Do NOT use for non-Python projects.