What you get

  • One offline HTML graph with search, type filters, focus and zoom
  • An evidence panel with file, line and quote for every claim
  • The JSON dataset, checked for broken ids and missing sources
  • A Markdown folder scanner that maps notes, links and tags
  • Optional cover image from PopNow, priced before it runs

Good for

  • Project handovers and onboarding packs
  • Research notes, wikis and decision logs
  • Audit or due-diligence folders that need traceable claims

Not for

  • Live data that changes every minute
  • Questions that need totals, trends or charts
  • Thousands of entities in one view

How to use it

In Claude or ChatGPT

Use the PopNow skill "Knowledge Graph Explorer" (knowledge-graph-explorer) for this: Map the people, decisions and open risks in ./handover so the new lead can see who decided what and why.

In a coding agent

/popnow-knowledge-graph-explorer Map the people, decisions and open risks in ./handover so the new lead can see who decided what and why.

Then try

Add the vendor contracts folder and mark anything inferred.

Hide document nodes by default and give me the list of unsourced claims.

What to have ready

  • The folder or records to include and what to leave out
  • Three questions the graph should answer
  • Entity types and relation verbs, or approval to propose them

Similar skills

View all →

Questions

Where do the entities and links come from?

From your files. The skill reads them as text, records the file and line for every entity and relation, and marks anything it inferred rather than read. A scanner handles Markdown links and tags automatically.

Does it need a server or a graph database?

No. The result is one HTML file with the data inside. It opens from disk and makes no network requests.

How big can the graph get?

Up to about 150 entities reads comfortably and up to about 400 still works. Larger archives are summarised or split into several pages.

What does PopNow do in this skill?

Only an optional cover or share image for the page that links to the graph, after you approve the price. The graph itself is always built from your data.