Synchronizes project documentation with code changes to maintain accurate context for AI assistants.
The LLM Documentation Manager skill streamlines the maintenance of project-specific documentation optimized for Claude and other large language models. It enforces a dual-structure system comprising a high-level map (CLAUDE.md) and detailed, concept-specific guides (.llmdocs/), ensuring that documentation remains concise, accurate, and perfectly synced with the current state of the code. By automating the identification of changes and proposing specific updates based on git diffs and session history, it eliminates documentation rot and helps Claude provide more relevant, context-aware assistance without manual overhead.
Key Features
01Automatic context gathering from session history and git diffs
02Incremental updates to existing documentation to prevent rewriting from scratch
03Dual-layer documentation architecture for high-level and deep-dive context
040 GitHub stars
05Enforced length constraints and formatting optimized for LLM readability
06Smart cross-referencing between project maps and specialized concept docs
Use Cases
01Onboarding AI agents to complex codebases with structured, living guides
02Updating project overview and architecture docs after implementing major features
03Maintaining an accurate directory of domain terms and business logic as code evolves
