wiki-export reads every page in a vault, builds a node and edge list from frontmatter, wikilinks and typed relationships, and writes the graph in several formats so external tools can use it.
One run produces graph.json, graph.graphml, cypher.txt, postgres.sql and a self-contained graph.html, grouped into communities by tag. An OKF bundle with full page bodies is available on request, and project and visibility filters narrow what is exported.
What You Get
- graph.json in NetworkX node-link format for scripts and graph libraries
- graph.graphml for Gephi, yEd and Cytoscape
- cypher.txt with MERGE statements for Neo4j
- postgres.sql with two tables and upsert statements
- graph.html, an interactive view colored by community and sized by link count
- An optional OKF bundle that keeps full page bodies for round trips through wiki-import
Common Use Cases
- Opening a vault’s link structure in Gephi or Neo4j for analysis
- Loading wiki pages and edges into Postgres for custom queries
- Sharing a public graph of a project while leaving out internal and private pages
What Makes It Unique
It covers five graph formats and a markdown bundle from one node and edge list, and it keeps typed relationships such as contradicts as named edge types in every format. Project filters match by folder or tag, and a public-export phrase drops pages tagged internal or private along with their edges. Re-running overwrites the outputs, and the notes advise gitignoring the derived folder.
Under The Hood
The skill is instructions for the agent. It reads the vault’s markdown files and writes the export files into a wiki-export folder inside the vault. The generated graph.html loads the vis.js library from a CDN when you open it in a browser, and the Cypher and SQL files are meant for you to load yourself with cypher-shell or psql. The skill does not run those loaders, fetch web pages or call remote APIs.