Understand Knowledge

Features

• Detects Karpathy-pattern LLM wikis
• Extracts entities, relationships, and topic clusters
• Generates an interactive knowledge graph dashboard
• Supports wikilink syntax and category-based organization
• Outputs graph data to local files for further use

Use Cases

• Exploring large LLM-generated knowledge bases
• Understanding the structure and connections within a wiki
• Identifying implicit relationships and claims discovered by LLM agents
• Onboarding new team members to a wiki knowledge base

Non-Goals

• Analyzing arbitrary markdown files not conforming to the Karpathy pattern
• Parsing binary files (e.g., PDFs) from raw source directories
• Providing a real-time collaborative editing interface for the graph

Workflow

1. Determine the target wiki directory
2. Execute the `parse-knowledge-base.py` script for deterministic scanning
3. Dispatch `article-analyzer` subagents to extract implicit knowledge
4. Execute the `merge-knowledge-graph.py` script to assemble the graph
5. Validate and save the final knowledge graph
6. Report summary and auto-trigger the dashboard

Practices

• Knowledge Graph Generation
• LLM-based Analysis
• Data Parsing and Merging
• Documentation Analysis

Prerequisites

• Python 3
• A directory conforming to the Karpathy-pattern LLM wiki structure