ワンクリックでManusで任意のスキルを実行

始める

knowledge-graph-generate

Analyze multiple documents to extract entities and relationships, generating a knowledge graph structure (RDF/Turtle).

Manusで実行

概要

Analyze multiple documents to extract entities and relationships, generating a knowledge graph structure (RDF/Turtle).

インストールコマンド

npx skills add https://github.com/dandye/information-architecture --skill knowledge-graph-generate

このコマンドをClaude Codeにコピー＆ペーストしてスキルをインストール

ソース

dandye/information-architecture

スター1

フォーク0

更新日2026年2月6日 23:03

ファイルエクスプローラー

2 ファイル

SKILL.md

readonly

name	knowledge-graph-generate
description	Analyze multiple documents to extract entities and relationships, generating a knowledge graph structure (RDF/Turtle).
required_roles	{"scribe":"roles/scribe.editor"}
personas	["information-architect","data-scientist","knowledge-manager"]

Knowledge Graph Generator Skill

Analyze a set of documents to identify named entities and their inter-relationships, constructing a knowledge graph to visualize and query the information structure.

Inputs

SOURCE_PATH - Directory or file list to analyze.
OUTPUT_FORMAT - (Optional) "turtle", "rdf/xml", "json-ld" (default: "turtle").
DEPTH - (Optional) Recursion depth for directory traversal (default: 1).

Workflow

Ingest: Read documents from SOURCE_PATH.
Extract: Perform Named Entity Recognition (NER) to find People, Organizations, Locations, Concepts.
Relate: Identify relationship verbs connecting entities (e.g., "authored by", "located in").
Serialize: Generate the graph in the requested OUTPUT_FORMAT (e.g., Turtle syntax).

Output

A serialized graph file or block representing the extracted knowledge.

このリポジトリの他の Skills

同じリポジトリ

named-entity-normalization

dandye/information-architecture

Normalize named entities and concept variants into canonical forms using controlled vocabularies or standard taxonomies.

2026-05-231

vocabulary-overlap-analysis

dandye/information-architecture

Analyze vocabulary overlap and identify named entities unique to domain/document corpora.

2026-05-231

document-tree-generate

dandye/information-architecture

Transform lengthy documents into a semantic tree structure. It extracts sections, summaries, and hierarchies optimized for use with Large Language Models (LLMs).

2026-04-091

ontology-define

dandye/information-architecture

Design formal ontologies using OWL/RDFS. Defines classes, properties, and relationships for complex semantic modeling.

2026-02-061

content-inventory

dandye/information-architecture

Systematic cataloging of information assets. Creates comprehensive inventories or card sorting materials from content.

2026-02-061

json-ld-generate

dandye/information-architecture

Generate JSON-LD structured data for web content. Maps content to Schema.org types to improve search engine understanding and rich result eligibility.

2026-02-061

ソース

dandye

dandye/information-architecture

GitHub リポジトリを開く Creator のリポジトリを見る

インストールコマンド

ダウンロード

Manusで実行

役立つ用途SOC

ソフトウェア開発者コンピュータ・数学職15-1252L4

name	knowledge-graph-generate
description	Analyze multiple documents to extract entities and relationships, generating a knowledge graph structure (RDF/Turtle).
required_roles	{"scribe":"roles/scribe.editor"}
personas	["information-architect","data-scientist","knowledge-manager"]

Knowledge Graph Generator Skill

Analyze a set of documents to identify named entities and their inter-relationships, constructing a knowledge graph to visualize and query the information structure.

Inputs

SOURCE_PATH - Directory or file list to analyze.
OUTPUT_FORMAT - (Optional) "turtle", "rdf/xml", "json-ld" (default: "turtle").
DEPTH - (Optional) Recursion depth for directory traversal (default: 1).

Workflow

Ingest: Read documents from SOURCE_PATH.
Extract: Perform Named Entity Recognition (NER) to find People, Organizations, Locations, Concepts.
Relate: Identify relationship verbs connecting entities (e.g., "authored by", "located in").
Serialize: Generate the graph in the requested OUTPUT_FORMAT (e.g., Turtle syntax).

Output

A serialized graph file or block representing the extracted knowledge.