extract-knowledge-graph

What problem does it solve?

This Skill converts a captured markdown documentation corpus into structured, queryable knowledge by extracting concepts, APIs, tools, entities, and the relationships between them.

Core Features & Use Cases

• Candidate extraction and semantic structuring: Detects likely concepts, API references, tool mentions, and entities from each captured page.
• Canonicalization with stable IDs: Collapses aliases into unified nodes and reuses existing indexed IDs when available, otherwise allocates new ULIDs.
• Relation extraction and graph emission: Produces relationship edges (e.g., extracted-from, defines, exposes, uses/part-of) and emits both JSON and Mermaid graph outputs.
• Optional repo-backed enrichment: Enriches API extraction from cloned source repositories (notably Rust via rustdoc JSON) when available.

Quick Start

Use the extract-knowledge-graph pipeline after scrape-firecrawl when you need structured semantic extraction beyond link graphs, such as turning a topic’s scraped markdown corpus into a Mermaid and JSON knowledge graph for that topic and tool.