skills.rest
ulebule
ulebuleCOMMUNITY
Explore all skills by this author →
GitHub

graphify

Converts folders of code, docs, papers, and media into a queryable knowledge graph with community detection.

Updated Aug 26, 2026
Search Tags:#knowledge graph#obsidian#codebase analysis#neo4j#graphrag#community detection#ast extraction
One-click install
npx skills add https://github.com/ulebule/claude-skills --skill graphify-ulebule
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: graphify
Source: https://github.com/ulebule/claude-skills/tree/main/plugins/graphify/skills/graphify
Command: npx skills add https://github.com/ulebule/claude-skills --skill graphify-ulebule

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires graphifyy, networkx, and includes references (resource) components.

What problem does it solve? Understanding a large codebase or mixed document corpus requires reading hundreds of files; graphify turns any folder into a persistent knowledge graph so questions about architecture, file relationships, and concepts are answered by graph traversal instead of manual searching. ## Core Features & Use Cases - Knowledge graph construction: Extracts entities and relationships from code (AST), docs, papers, images, and transcribed video, with an honest EXTRACTED/INFERRED/AMBIGUOUS audit trail. - Analysis and outputs: Community detection, god-node and surprising-connection analysis, plus interactive HTML, GraphRAG-ready JSON, GraphML, SVG, Neo4j/FalkorDB exports, Obsidian vaults, and a plain-language GRAPH_REPORT.md. - Query tools: BFS/DFS traversal queries, shortest-path between concepts, and plain-language node explanations, with incremental --update and --watch for keeping the graph current. - Use Case: Point it at a cloned GitHub repo (or several repos merged into one cross-repo graph), then ask "How does the auth flow reach the database?" and get an answer traced through the graph with source citations. ## Quick Start Run /graphify on the current project folder to build the knowledge graph, then ask a natural-language question about the codebase.

Frequently Asked Questions about graphify

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a knowledge graph from a codebase?▼

Run /graphify on a local path or GitHub URL. It detects files, runs AST extraction on code and parallel LLM subagents on docs and images, then clusters the merged graph and writes graph.json, an interactive HTML view, and a GRAPH_REPORT.md.

How do I query an existing graphify knowledge graph?▼

Once graphify-out/graph.json exists, run graphify query "<question>" for BFS traversal, add --dfs to trace a path, or use graphify path "A" "B" and graphify explain "X". Queries are expanded against the graph's actual vocabulary before matching.

Can graphify handle GitHub repos and multiple repositories?▼

Yes. Pass one or more GitHub URLs and graphify clones them into ~/.graphify/repos, builds each graph, and merges them with graphify merge-graphs into a cross-repo graph where every node carries a repo attribute.

Does graphify support video and audio files?▼

Yes. Detected video and audio files are transcribed with Whisper (model selectable via --whisper-model) using a domain hint derived from the corpus, and the transcripts are then treated as documents during semantic extraction.

How do I update the graph after code changes without re-extracting everything?▼

Use /graphify --update for incremental re-extraction of only new or changed files, with code-only changes skipping LLM extraction entirely. Alternatively, --watch monitors a folder and rebuilds automatically, and a post-commit hook can be installed.

What export formats does graphify support?▼

Graphify exports interactive HTML (default), GraphRAG-ready JSON, SVG, GraphML, an Obsidian vault, an agent-crawlable wiki, and Cypher for Neo4j or FalkorDB with direct push options. It can also run as an MCP stdio server for live agent queries.

Related Skills

Explore More Software Engineering Skills→
zarazhangruizarazhangrui
5.5k

codebase-to-course

Generate an interactive single-page HTML course from any codebase.

Community
Advanced
FalkorDBFalkorDB
22

Migrate Neo4j to FalkorDB

Extracts Neo4j nodes and relationships into CSV files for FalkorDB import.

Official
Intermediate
YouneslahlaliYouneslahlali

extract

Extract reusable UI components and design tokens into a design system.

Community
Advanced
zilliztechzilliztech
3

semantic-search

Convert text into embeddings and retrieve semantically similar items from a vector store.

Official
Intermediate