Agent skill · volcengine
knowledge-graph
Compile documents, notes, web content, transcripts, research materials, or code repositories into an evidence-grounded, visualization-ready knowledge graph with semantically typed entity nodes, statement-level provenance, and typed relationship edges. Use with ov compile to create or incrementally refresh `entities/*.md` node artifacts and a root `relations.jsonl` edge file for people, organizations, groups, animals, places, products, projects, systems, services, documents, events, and other identifiable things.
What it needs
About 7k tokens when loaded.
What this skill does
Knowledge Graph Goal Turn the supplied sources into a durable graph that agents can traverse by entity, type, and relationship. Build the graph as this artifact tree: Keep sources read-only. Follow the task instruction for scope, language, audience, and depth; otherwise use the dominant language of the sources. Ground every node, material claim, and edge in the supplied sources. Treat sources as provenance rather than domain nodes unless a source is itself a named subject in the domain. Entity nodes Create a node for a named thing with a stable identity or boundary, such as a person, organization, group, animal, place, product, project, system, service, module, dataset, standard, document, artifact, or named event. Store each node at entities/<entity-id>.md with this structure: Use type: entity as the artifact kind and entitytype as the entity's semantic class. Choose a stable lowercase snakecase value such as person, organization, group, animal, place, product, project, system, service, module, dataset, standard, document, artifact, or event. Reuse an established type vocabulary when refreshing a graph; introduce a narrower type only when it materially improves filtering or visualization. Write description as one stable, context-independent identity sentence. Do not put a single episode's actions, a temporary state, a source-specific opinion, or an extraction history in the description. Record exact source references that establish the entity's identity and description in the non-empty sources list. This node-level list does not replace claim-specific evidence. Follow the frontmatter with concise Markdown only when it adds useful context beyond the fingerprint. Describe applicable stable attributes, responsibilities, interfaces, boundaries, or source-scoped context. Use level-2 headings such as ## Overview or their equivalent in the output language, and omit empty sections. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills volcengine/knowledge-graph