apifox_skill

Overview

This skill provides a command-line tool to read, cache, search, and export the ABC Medical Cloud OpenAPI documentation (about 4–5 MB, ~4200 endpoints). It exposes commands to list modules, search paths, fetch full endpoint definitions with automatic $ref resolution, and generate summary exports. The tool caches the OpenAPI document locally for fast, offline queries and supports manual refresh when you need the latest spec.

How this skill works

The skill calls the Apifox API using an access token to fetch the project OpenAPI export and stores the JSON in a local cache directory. It builds an in-memory index so searches (by module, path, method, or keyword) return results in milliseconds. Commands can retrieve a single path’s full definition and automatically resolve $ref references, or skip ref resolution for faster responses.

When to use it

• Investigate or review a specific API endpoint definition including request/response schema.
• Search for endpoints by keyword, module, path pattern, or HTTP method across the entire ABC API surface.
• Generate an API summary or Markdown/JSON export for documentation or handoff.
• Inspect module-level metrics and overall endpoint statistics during architecture or QA work.
• Work offline after the initial fetch using the cached OpenAPI document.

Best practices

• Set APIFOX_ACCESS_TOKEN and APIFOX_PROJECT_ID in your shell environment before first use.
• Run read_oas or refresh_oas once to populate the local cache before heavy querying.
• Use search_paths with module and method filters to narrow results and reduce noise.
• Prefer get_path with include_refs=true when you need complete, de-referenced schemas for implementation.
• Use cache_status and clear_cache when you suspect stale data or need to reload the source.

Example use cases

• Developer examining the POST /api/global-auth/login/sms definition to implement SMS login in a client app.
• QA engineer searching for all endpoints that accept multipart/form-data to validate file upload behavior.
• Technical writer exporting module summaries to Markdown for an internal API reference site.
• Architect generating statistics to assess API surface area and prioritize refactor work.
• Support engineer quickly finding error response schemas for troubleshooting production issues.