document-relate

name	document-relate
description	Create or query directed relations between documents as graph edges.
inputs	[{"name":"document_id","type":"string","required":true,"description":"UUID of the source Document."},{"name":"relation_type","type":"string","required":true,"description":"Relation label — e.g. \"supersedes\", \"references\", \"derived-from\", \"related-to\"."},{"name":"target_document_id","type":"string","required":false,"description":"UUID of the target Document. Required for \"create\"; omit to query existing relations."},{"name":"action","type":"string","required":false,"description":"One of \"create\" or \"query\". Defaults to \"query\"."}]
outputs	[{"name":"relations","type":"array","description":"List of {relation_id, relation_type, target_document_id, target_title} for the source document."},{"name":"relation_id","type":"string","description":"UUID of the newly created relation row; populated for \"create\" action only."}]
tools	["db-query"]

document-relate

When to use

Invoke this skill when the user says a document supersedes, references, or is derived from another, or when they ask which documents are related to a given one. Use it to build the knowledge graph before running document-search over a topic cluster.

How it works

• For "create": validate that both Document rows exist via db-query, then insert a relation edge into the document relations join table linked to both Document UUIDs.
• For "query": fetch all outbound and inbound relation edges for document_id from the relations table; enrich each with the related Document.title via db-query.
• Enforce that no duplicate (source, type, target) triplet is inserted.

Audit

Emits one ExecutionHistory row with event_type=DOCUMENT_EDITED and actor_type=agent for "create". The "query" action is read-only and emits no row.