Converts folders of code, docs, papers, and images into a clustered knowledge graph with HTML, JSON, and audit report outputs.
npx skills add https://github.com/unnxt30/skills --skill graphify-unnxt30
Please help me install this Agent Skill. Skill: graphify Source: https://github.com/unnxt30/skills/tree/main/graphify Command: npx skills add https://github.com/unnxt30/skills --skill graphify-unnxt30
High-intent search queries and answers about installing and using this skill.
Run /graphify on the target path. It detects file types, extracts code structure via AST parsing, dispatches parallel subagents for docs and images, clusters the result into communities, and outputs interactive HTML plus GraphRAG-ready JSON.
graphify handles code files, Markdown and text docs, PDF papers, images, and video or audio files. Video and audio are transcribed with Whisper first, then treated as documents during semantic extraction.
Yes. Use --neo4j to generate a Cypher import file or --neo4j-push to write directly to a running Neo4j instance. The --obsidian flag generates a vault with one note per node plus a canvas file for community layout.
If a corpus exceeds 2 million words or 200 files, graphify shows the top subdirectories and asks which subfolder to process. Extraction results are cached, so incremental runs with --update only reprocess new or changed files.
Every edge carries a confidence label: EXTRACTED for relationships explicit in the source, INFERRED for reasoned connections with a 0.4-0.9 confidence score, and AMBIGUOUS for uncertain ones. This audit trail separates found facts from model inference.
HTML visualization is skipped for graphs over 5000 nodes because they become unusable in a browser. For very large graphs, use the Obsidian vault export or query the graph.json output programmatically instead.
yxy23Process Zotero papers into structured vault notes with highlights and marginal notes.
hiddink-aiQuery project wiki documentation to answer technical questions with inline citations.
jonnabioIndex and retrieve documents using semantic embeddings and metadata.
zMynxxBuilds a queryable knowledge graph from code, docs, papers, images, and video.
