Frontend Repo Business Skill Generator

Problem

Frontend repository logic, API calls, and business rules are often scattered across routes, components, handler functions, and third-party component props. Manually turning that into an operation manual can easily produce an API dump instead of a business guide, while missing key preconditions, parameter sources, confirmation steps, and error handling. code-skill aims to extract executable business Skills from real repository workflows, so an AI can ask, check, and perform actions in a way that matches actual product behavior.

How It Works

• CLI-first grounding: It expects an existing business CLI and reads the real command and parameter surface via --help; if no CLI exists, it first requires a usable CLI before proceeding.
• Repository scanning: It starts with route files to identify pages, modules, and hierarchy, then searches event handlers such as handle* and on* to map operation entry points.
• Selective reading: It focuses on parameter sources, guard conditions, call order, user confirmation points, and error handling, while skipping JSX, CSS, and purely visual state.
• Capability alignment: Only steps that can be mapped to real CLI commands are included; unmapped actions are marked as unsupported.
• Skill output: It generates Markdown Skills by type—query, operation, analysis, or workflow—with trigger conditions, parameter collection, execution steps, and error handling.

Boundaries

It is best for frontend repositories that already expose business actions through a CLI. It does not replace API documentation or invent capabilities the CLI cannot support; field semantics must be validated against code and page data before being written into a Skill.