Knowledge Graph Mapper

The Problem

When engineers maintain Markdown or TXT documents, people, organizations, technologies, concepts, and recurring terms are often scattered across files. Manual mapping is easy to miss, and plain full-text search only locates strings without showing how entities relate. knowledge-mapper targets this lightweight knowledge-organization problem: turning loose documents into queryable and exportable entity-relationship structures, without pretending to perform deep semantic inference.

How It Works

• Document parsing: reads Markdown and TXT files as knowledge-base input.
• Entity recognition: extracts typed entities such as PERSON, ORG, TECH, CONCEPT, and TERM using rule- and keyword-based heuristics.
• Relation discovery: records entity co-occurrences to show which people, organizations, concepts, or technologies appear together.
• Knowledge querying: supports searching entities and documents, viewing document lists, entities, relations, and statistics.
• Knowledge visualization: exports text, JSON, and GraphViz DOT for later graph rendering.

Scope and Caveats

This skill is suitable for early-stage knowledge mapping and keyword-level organization, not for precise semantic understanding, entity disambiguation, or relation inference. Current extraction is rule-driven, so complex terminology, synonyms, and cross-document coreference may be missed.