Tiling Trees MCP Server

Instructions

Implementation Reference

• src/research-tree.ts:427-466 (handler)

  Core handler function getTopLeaves in ResearchTreeManager class that filters evaluated leaf tiles, computes scores based on specified criteria (impact, feasibility, uniqueness, or combined average), sorts descending, and returns top N results.
  * Get top-rated leaf tiles based on evaluation criteria */ getTopLeaves( criteria: "impact" | "feasibility" | "uniqueness" | "combined", limit: number = 10, treeId?: string ): Tile[] { const leaves = this.getLeafTiles(treeId).filter((tile) => tile.evaluation); const scored = leaves.map((tile) => { let score = 0; const evaluation = tile.evaluation!; switch (criteria) { case "impact": score = evaluation.impact || 0; break; case "feasibility": score = evaluation.feasibility || 0; break; case "uniqueness": score = evaluation.uniqueness || 0; break; case "combined": score = ((evaluation.impact || 0) + (evaluation.feasibility || 0) + (evaluation.uniqueness || 0)) / 3; break; } return { tile, score }; }); return scored .sort((a, b) => b.score - a.score) .slice(0, limit) .map((item) => item.tile); }
• src/index.ts:280-302 (schema)

  Input schema definition for the get_top_leaves tool, specifying required 'criteria' parameter and optional 'limit' and 'treeId'.
  { name: "get_top_leaves", description: "Get the highest-rated leaf tiles based on evaluation criteria", inputSchema: { type: "object", properties: { criteria: { type: "string", enum: ["impact", "feasibility", "uniqueness", "combined"], description: "Criteria to sort by", }, limit: { type: "number", description: "Number of results to return (default: 10)", }, treeId: { type: "string", description: "Optional tree ID to filter by", }, }, required: ["criteria"], }, },
• src/index.ts:572-586 (handler)

  Dispatch handler in the main CallToolRequestSchema switch statement that invokes treeManager.getTopLeaves with parsed arguments and formats the JSON response.
  case "get_top_leaves": { const result = treeManager.getTopLeaves( args.criteria as any, args.limit as number | undefined, args.treeId as string | undefined ); return { content: [ { type: "text", text: JSON.stringify(result, null, 2), }, ], }; }
• src/index.ts:27-390 (registration)

  The get_top_leaves tool is included in the TOOLS array registered with server.setRequestHandler(ListToolsRequestSchema) to advertise available tools.
  const TOOLS: Tool[] = [ { name: "create_tree", description: "Create a new tiling tree to explore a problem/challenge. The tree starts with a root tile representing the complete solution space, which you'll then split recursively using MECE (Mutually Exclusive, Collectively Exhaustive) principles.", inputSchema: { type: "object", properties: { name: { type: "string", description: "Name for this tiling tree", }, problemStatement: { type: "string", description: "The problem or challenge to explore (e.g., 'How can we reduce carbon emissions in transportation?')", }, }, required: ["name", "problemStatement"], }, }, { name: "split_tile", description: "Split a tile into MECE (Mutually Exclusive, Collectively Exhaustive) subsets using a specific attribute/dimension. This is the core operation of the tiling trees method - partitioning the solution space systematically. Use physics/math-oriented splits when possible.", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile to split", }, splitAttribute: { type: "string", description: "The attribute/dimension used to split (e.g., 'energy source', 'scale', 'physical mechanism', 'timeframe')", }, splitRationale: { type: "string", description: "Why this attribute was chosen for splitting", }, subsets: { type: "array", items: { type: "object", properties: { title: { type: "string" }, description: { type: "string", description: "Precise definition of this subset to ensure no overlap with siblings", }, isLeaf: { type: "boolean", description: "True if this is a concrete idea/project (leaf node)", }, }, required: ["title", "description"], }, description: "The mutually exclusive and collectively exhaustive subsets", }, }, required: ["tileId", "splitAttribute", "splitRationale", "subsets"], }, }, { name: "add_tiles_to_split", description: "Add additional tiles to an existing split (when you realize a category was missed). This invalidates the MECE validation and requires re-verification.", inputSchema: { type: "object", properties: { parentId: { type: "string", description: "ID of the parent tile", }, newTiles: { type: "array", items: { type: "object", properties: { title: { type: "string" }, description: { type: "string" }, isLeaf: { type: "boolean" }, }, required: ["title", "description"], }, description: "New tiles to add to the split", }, }, required: ["parentId", "newTiles"], }, }, { name: "mark_mece", description: "Mark a split as validated for MECE (Mutually Exclusive, Collectively Exhaustive) properties. Verify that the children completely cover the parent space with no overlaps.", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile whose split to validate", }, isMECE: { type: "boolean", description: "Whether the split is truly MECE", }, coverageNotes: { type: "string", description: "Notes on the completeness and exclusivity of the split", }, }, required: ["tileId", "isMECE"], }, }, { name: "evaluate_tile", description: "Evaluate a leaf tile (concrete idea/project) on impact, feasibility, and uniqueness. Include any calculations or pilot studies performed.", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile to evaluate", }, impact: { type: "number", description: "Impact rating (1-10 scale)", minimum: 1, maximum: 10, }, feasibility: { type: "number", description: "Feasibility rating (1-10 scale)", minimum: 1, maximum: 10, }, uniqueness: { type: "number", description: "Uniqueness rating (1-10 scale)", minimum: 1, maximum: 10, }, timeframe: { type: "string", description: "Expected timeframe (e.g., '1-2 years', '5-10 years')", }, notes: { type: "string", description: "Additional evaluation notes", }, calculationsOrPilots: { type: "string", description: "Calculations or pilot studies performed to evaluate this idea", }, }, required: ["tileId"], }, }, { name: "update_tile", description: "Update a tile's information (title, description, split attributes, etc.)", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile to update", }, title: { type: "string", description: "New title", }, description: { type: "string", description: "New description (precise definition)", }, splitAttribute: { type: "string", description: "Updated split attribute", }, splitRationale: { type: "string", description: "Updated split rationale", }, isLeaf: { type: "boolean", description: "Mark as leaf node", }, }, required: ["tileId"], }, }, { name: "get_trees", description: "Get all tiling trees", inputSchema: { type: "object", properties: {}, }, }, { name: "get_tile", description: "Get details of a specific tile", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile", }, }, required: ["tileId"], }, }, { name: "explore_path", description: "Explore the tree structure from a specific tile, showing the hierarchical breakdown", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile to explore from", }, depth: { type: "number", description: "How many levels deep to explore (default: 10)", }, }, required: ["tileId"], }, }, { name: "get_leaf_tiles", description: "Get all leaf tiles (concrete ideas/projects) from a tree", inputSchema: { type: "object", properties: { treeId: { type: "string", description: "Optional tree ID to filter by", }, }, }, }, { name: "get_unexplored_tiles", description: "Get tiles that haven't been split yet - these are gaps in your solution space exploration", inputSchema: { type: "object", properties: { treeId: { type: "string", description: "Optional tree ID to filter by", }, }, }, }, { name: "get_top_leaves", description: "Get the highest-rated leaf tiles based on evaluation criteria", inputSchema: { type: "object", properties: { criteria: { type: "string", enum: ["impact", "feasibility", "uniqueness", "combined"], description: "Criteria to sort by", }, limit: { type: "number", description: "Number of results to return (default: 10)", }, treeId: { type: "string", description: "Optional tree ID to filter by", }, }, required: ["criteria"], }, }, { name: "search_tiles", description: "Search for tiles by content", inputSchema: { type: "object", properties: { query: { type: "string", description: "Search query", }, treeId: { type: "string", description: "Optional tree ID to filter by", }, }, required: ["query"], }, }, { name: "get_coverage_analysis", description: "Analyze the completeness of solution space exploration for a tree. Shows unexplored branches, unvalidated splits, and suggestions for next steps.", inputSchema: { type: "object", properties: { treeId: { type: "string", description: "ID of the tree to analyze", }, }, required: ["treeId"], }, }, { name: "get_statistics", description: "Get overall statistics about all tiling trees", inputSchema: { type: "object", properties: {}, }, }, { name: "export_tree", description: "Export a tiling tree in various formats for visualization or documentation", inputSchema: { type: "object", properties: { treeId: { type: "string", description: "ID of the tree to export", }, format: { type: "string", enum: ["json", "markdown", "mermaid", "dot"], description: "Export format", }, }, required: ["treeId", "format"], }, }, { name: "validate_split_quality", description: "Validate split quality and detect common antipatterns (vague language, catch-all buckets, mixed dimensions, retroactive splitting, incomplete coverage). Returns a detailed quality report with issues and recommendations.", inputSchema: { type: "object", properties: { tileId: { type: "string", description: "ID of the tile whose split to validate", }, }, required: ["tileId"], }, }, { name: "get_tree_validation_report", description: "Get validation report for all splits in a tree. Identifies antipatterns and provides an overall quality score. Use this after building a tree to check for common failure modes.", inputSchema: { type: "object", properties: { treeId: { type: "string", description: "ID of the tree to validate", }, }, required: ["treeId"], }, }, ];