AGI MCP Server

Instructions

Implementation Reference

• src/memory-manager.js:907-928 (handler)

  Main handler implementation that executes SQL query to find clusters similar to a given cluster using vector cosine similarity (pgvector <=> operator). Takes clusterId and threshold parameters, returns array of similar clusters with similarity scores.

  async findSimilarClusters(clusterId, threshold = 0.7) {
    try {
      const embeddingVector = `[${Array(1536).fill(0).join(',')}]`;
      
      const results = await this.db.execute(sql`
        SELECT 
          mc2.*,
          1 - (mc1.centroid_embedding <=> mc2.centroid_embedding) as similarity
        FROM memory_clusters mc1
        CROSS JOIN memory_clusters mc2
        WHERE mc1.id = ${clusterId}
          AND mc2.id != ${clusterId}
          AND 1 - (mc1.centroid_embedding <=> mc2.centroid_embedding) >= ${threshold}
        ORDER BY similarity DESC
      `);
      
      return results.rows || [];
    } catch (error) {
      console.warn('Similar clusters query failed:', error.message);
      return [];
    }
  }

• src/tools/memory-tools.js:362-379 (schema)

  Tool schema definition in the tools registry. Defines input parameters: cluster_id (required string) and threshold (optional number with default 0.7). Describes the tool as finding clusters similar to a given cluster.

    name: "find_similar_clusters",
    description: "Find clusters similar to a given cluster",
    inputSchema: {
      type: "object",
      properties: {
        cluster_id: {
          type: "string",
          description: "UUID of the reference cluster"
        },
        threshold: {
          type: "number",
          description: "Minimum similarity threshold",
          default: 0.7
        }
      },
      required: ["cluster_id"]
    }
  },

• mcp.js:387-405 (registration)

  Tool registration in the MCP server's tools list. Contains the tool name, description, and input schema definition with cluster_id and threshold parameters.

  {
    name: "find_similar_clusters",
    description: "Find clusters similar to a given cluster",
    inputSchema: {
      type: "object",
      properties: {
        cluster_id: {
          type: "string",
          description: "UUID of the reference cluster"
        },
        threshold: {
          type: "number",
          description: "Minimum similarity threshold",
          default: 0.7
        }
      },
      required: ["cluster_id"]
    }
  },

• mcp.js:649-654 (registration)

  Switch case handler that routes tool invocations to the memoryManager.findSimilarClusters method, passing cluster_id and threshold (defaulting to 0.7), and returns results as formatted JSON text.

  case "find_similar_clusters":
    const similarClusters = await memoryManager.findSimilarClusters(
      args.cluster_id,
      args.threshold || 0.7
    );
    return { content: [{ type: "text", text: JSON.stringify(similarClusters, null, 2) }] };