dev-llm-docgen

What problem does it solve?

This Skill removes the friction of writing first-pass Doxygen documentation for C, C++, and CUDA functions by generating a structured comment draft from a symbol name.

Core Features & Use Cases

• Docblock Drafting: Produces a Doxygen-style block with @brief, @param, and @return content for a named function.
• Local LLM Workflow: Uses a local Ollama-backed toolchain to generate documentation without sending code to an external service.
• Safe Review Loop: Verifies the target symbol exists, shows the draft to the user, and lets the user choose whether to insert, copy, or regenerate it.
• CUDA-Aware Guidance: Adjusts parameter wording for GPU code so descriptions better fit device and host memory concepts.
• Use Case: A developer has an undocumented kernel or helper function and wants a polished docblock draft before committing it to the codebase.

Quick Start

Ask the skill to generate a Doxygen docblock for the named function in your C, C++, or CUDA file and then choose whether to insert, copy, or regenerate the draft.