Synchronizes codebase changes with project documentation to maintain accurate context for AI assistants.
The update-llm-docs skill automates the maintenance of project-specific documentation designed to optimize Large Language Model performance. It manages a two-tiered system: a high-level CLAUDE.md 'map' for immediate context and a detailed .llmdocs/ 'territory' for deep-dives into architecture, APIs, and data models. By analyzing session history and git diffs, the skill ensures that the AI's understanding of the codebase remains current, preventing hallucinations and errors caused by stale documentation while adhering to strict brevity and accuracy rules.
主な機能
01Standardized templates for architecture, data models, and API documentation
02Smart identification of required documentation updates via session history analysis
03Automated synchronization between code changes and CLAUDE.md overviews
040 GitHub stars
05Strict line-count enforcement to optimize AI token usage and focus
06Incremental updates to .llmdocs/ files to preserve existing knowledge
ユースケース
01Standardizing documentation across multiple projects to ensure consistent Claude performance
02Updating project overview and conventions after a major refactor or feature launch
03Maintaining a specialized documentation layer specifically for AI context without cluttering user docs
