What problem does it solve? Writing developer documentation that matches the actual system is slow and error-prone; docs drift from code, mix proposals with shipped behavior, and break links or structure. This Skill inspects the real code, tests, schemas, and interfaces before writing, then checks the finished Markdown with a bundled validator. ## Core Features & Use Cases - Behavior-grounded writing: Reads current code, tests, configuration, and command help before documenting, and labels anything that cannot be verified instead of inventing claims. - Full document-type coverage: Produces READMEs, quickstarts, tutorials, how-to guides, concept pages, API/SDK/CLI references, installation and migration guides, runbooks, release notes, and troubleshooting pages, each matched to the reader's job. - Bundled Markdown validator: A dependency-free Python script checks UTF-8 decoding, heading structure, fenced code blocks, local links, image alt text, and optional GitHub-style heading fragments. - Use Case: Ask it to build a documentation set for a CLI project; it inspects the command parser, writes an overview, quickstart, and CLI reference with verified commands, then runs the validator to confirm links and structure pass. ## Quick Start Ask the AI to create a README and quickstart for this project using the developer-documentation skill, verifying every documented command against the current code.
