This skill is designed for building and maintaining LLM-driven knowledge bases, specifically following the 'Karpathy-style' research workflow. It automates the process of evolving a research project by locating existing YAML outlines, prompting the user for new subjects, and optionally deploying a web search agent to discover additional relevant topics. By handling duplicate prevention and file updates automatically, it allows researchers and developers to scale their personal wikis or technical documentation with minimal manual overhead.
主な機能
01Parallel collection of items from user input and web searches
02Confirmation-based workflow for verified updates
03In-place modification of research outline files
0429 GitHub stars
05Automated discovery of outline.yaml files in the workspace
06Duplicate detection to ensure clean data merging
ユースケース
01Maintaining an LLM-powered Obsidian knowledge base
02Scaling a technical wiki with new sub-topics and research subjects
03Automating the discovery phase of deep-dive technical research