Document React components with AI
Give an agent a concrete reason to choose an existing component. Record its inputs, expected behaviour and a working example beside the code.
When you document React components with AI, the first draft can come straight from the component file. The ai-component-metadata Python helper reads a local TSX file and writes a separate TypeScript metadata file in the same directory. It recognises common source patterns, then leaves room for an explanation that only someone familiar with the interface can supply.
Consider a compact menu button. Its icon might be obvious from the source, while the expected focus position after closing the menu is a decision worth recording. Put that decision beside a working example so a colleague or an agent can check it when composing another screen.
What it does
- Identify required inputs and common states in a simple component.
- Explain where a component belongs within a larger form or menu.
- Provide a working composition from your own library.
- Describe a plausible misuse and name a better available component.
- Check keyboard interactions and screen reader expectations in a rendered example.
Install and use
- Choose a local React component written in TypeScript. Use Python for the starter file, or fill in the supplied template by hand.
- Give the agent a specific example such as a menu that returns focus to its trigger when it closes.
- Open the generated sibling file, replace the placeholder descriptions and test the example before sharing it.
FAQ
Source
Made by AI Component Metadata Contributors. Original skill: ai-component-metadata, published on GitHub under the MIT licence.
View on GitHub