Intelligence Tools

Tools for querying the research intelligence context assembly system. The intelligence query assembles a structured context bundle from hypotheses, experiments, the knowledge graph, and detected patterns.

Available Tools

intelligence_query

Query the research intelligence context — "What do we know about X?"

This tool assembles a comprehensive context bundle by searching across typed pages, the research graph, and pattern detections to answer a natural-language query.

{
  "name": "intelligence_query",
  "arguments": {
    "workspaceId": "ws_123",
    "query": "What do we know about API caching performance?",
    "spaceId": "space_456"
  }
}

Arguments:

Argument	Type	Required	Description
workspaceId	string	Yes	Workspace ID
query	string	Yes	Natural-language query
spaceId	string	No	Optional space filter

Response:

{
  "context": {
    "query": "What do we know about API caching performance?",
    "directHits": [
      {
        "id": "page_hyp_123",
        "title": "Caching reduces API latency by 50%",
        "pageType": "hypothesis"
      }
    ],
    "relatedWork": [
      {
        "id": "page_789",
        "title": "Caching Architecture Decisions",
        "pageType": "finding"
      }
    ],
    "timeline": [
      {
        "id": "page_exp_456",
        "title": "Redis cache benchmark",
        "updatedAt": "2024-01-15T14:30:00Z"
      }
    ],
    "currentState": {
      "validated": [
        { "id": "page_hyp_123", "title": "Caching reduces API latency by 50%" }
      ],
      "refuted": [],
      "testing": [],
      "open": []
    },
    "openQuestions": [
      {
        "id": "task_oq_123",
        "title": "What is the cache invalidation strategy?",
        "status": "todo",
        "priority": "medium",
        "labels": ["open-question"]
      }
    ],
    "contradictions": [
      {
        "from": "page_hyp_123",
        "to": "page_hyp_456",
        "type": "CONTRADICTS"
      }
    ],
    "experiments": [
      {
        "id": "page_exp_456",
        "title": "Redis cache benchmark",
        "pageType": "experiment"
      }
    ],
    "papers": []
  }
}

Context Bundle Structure

The intelligence query assembles context from multiple sources:

Section	Source	Description
directHits	Full-text search	Pages matching the query via tsquery
relatedWork	Knowledge graph	Pages connected via graph traversal (max depth 2)
timeline	Page metadata	All matched pages sorted by last update
currentState	Typed pages	Hypotheses grouped by status (validated, refuted, testing, open)
openQuestions	Tasks	Tasks labeled "open-question" or matching the query
contradictions	Memgraph	CONTRADICTS edges involving result pages
experiments	Typed pages	Experiments related to matched hypotheses
papers	Typed pages	Papers and formal publications

Use Cases

Agent Context Injection

Use intelligence_query to give AI agents full research context before they take action:

// Before an agent writes a recommendation
const context = await mcp.call("intelligence_query", {
  workspaceId: "ws_123",
  query: "database migration strategies"
});

// Agent now has hypotheses, evidence, and patterns to inform its response

Research Status Check

Quickly understand the state of knowledge in a domain:

const context = await mcp.call("intelligence_query", {
  workspaceId: "ws_123",
  query: "authentication security",
  spaceId: "space_security"
});

// See which hypotheses are validated, what's still open, any contradictions

Related

• Hypothesis Tools - Create and manage hypotheses
• Experiment Tools - Register experiments
• Pattern Tools - Pattern detection
• Research Intelligence Guide - Full system overview