docs-writer

What This Skill Does

The docs-writer skill is a specialized technical documentation agent designed to streamline the creation of API documentation, README files, and user guides. By leveraging an incremental generation workflow, it prevents token overflow and logical fragmentation, ensuring that documentation remains coherent and highly structured. It follows strict technical writing principles, prioritizing 'show-don't-tell' via code examples and maintaining clear, concise language suitable for developers and end-users.

Installation

To integrate this skill into your OpenClaw environment, use the following command: clawhub install openclaw/skills/skills/anton-abyzov/sw-docs-writer

Use Cases

• API Documentation: Generating comprehensive endpoint references including authentication, request structures, and error handling.
• README Maintenance: Standardizing project entry points with consistent installation steps, quick-start guides, and feature lists.
• Developer Onboarding: Creating structured tutorials that guide new engineers through complex system architectures.
• Technical Audits: Documenting system behavior by systematically walking through modules and functionality.

Example Prompts

1. "docs-writer, please start a new README for our authentication service, focusing on the OAuth2 integration workflow."
2. "I need to document our latest API release. Please generate an API reference section for the /v2/billing endpoint, including the 402 and 403 error codes."
3. "Could you create a 'Getting Started' user guide for the new dashboard, covering initial setup through to the first report generation?"

Tips & Limitations

• Incremental Workflow: Always allow the skill to proceed section-by-section. Attempting to force full document generation can result in truncated responses due to token limits.
• Frontmatter Importance: Ensure you keep the TL;DR frontmatter block in your documents, as it is critical for LLM context optimization.
• Verification: While the skill follows rigorous formatting, verify all technical specs (like endpoints or package names) against your actual codebase, as the agent relies on the context provided to it during the analysis phase.
• Scope: Keep your analysis phase requests specific to ensure the agent maintains high accuracy during the generation phase.