/graph-query

You can query a knowledge graph using natural language requests like "high priority tasks" or structured command-line parameters such as --status or --priority to filter and retrieve specific notes. This allows precise filtering of your JSON index without manual searching.

Finding orphaned notes and broken links in a knowledge base is done by running special graph queries that scan the JSON index to identify disconnected content and invalid references. This surfaces isolated documents and broken connections directly.

Searching for Architecture Decision Records (ADRs) with a specific status is done by querying the knowledge graph index using natural language phrases like "ADRs with status proposed". You can also use structured parameters like --type=Adr for precise filtering.

Yes, you can use natural language to search a JSON knowledge graph index by asking questions that filter notes by type, status, priority, or content. The command-line interface script interprets these queries to return matching records.

The graph query tool requires the knowledge base index to be stored in JSON format. It relies on the @anthropic-ai/sdk dependency to process natural language and structured parameter-based queries against this JSON data.

Using command-line parameters for graph queries limits you to filtering by predefined fields like type, status, and priority. For complex discovery tasks like mapping relationships, you must rely on the script's built-in special query functions rather than custom parameter combinations.