DeepWiki 文档生成 Skill

Generating structured technical documentation from source code requires analyzing the project type, selecting an organization mode, extracting chapter-relevant source files, and writing evidence-based content with source code line number references.

Evidence-based writing for software architecture overviews means mapping documentation chapters directly to extracted source files, ensuring technical claims are backed by verifiable code references rather than existing solely in chat context.

Creating Mermaid diagrams for technical documentation automatically involves generating Chinese-annotated charts for architecture and process flows, followed by quality checks to ensure structural completeness and SPEC format compliance.

Yes, you can enforce spec compliance for project documentation by applying SPEC.md language and format rules during generation, executing post-generation quality checks, and enforcing output paths relative to the AGENTS.md project root.

The best way to map source code files to documentation chapters is to determine the project type and organization mode first, then extract chapter-specific source files to populate required sections like purpose and scope with verifiable references.

Yes, you need an AGENTS.md file in the project root directory to establish the output target, alongside the docs/deepwiki/ path, ensuring the generated technical documentation maintains correct relative pathing and evidence chains.