@kazuph/mcp-taskmanager

Instructions

Implementation Reference

• index.ts:369-414 (handler)

  Core handler function that executes the tool logic: generates IDs, creates tasks, persists to JSON file, generates progress table.
  public async requestPlanning( originalRequest: string, tasks: { title: string; description: string }[], splitDetails?: string ) { await this.loadTasks(); this.requestCounter += 1; const requestId = `req-${this.requestCounter}`; const newTasks: Task[] = []; for (const taskDef of tasks) { this.taskCounter += 1; newTasks.push({ id: `task-${this.taskCounter}`, title: taskDef.title, description: taskDef.description, done: false, approved: false, completedDetails: "", }); } this.data.requests.push({ requestId, originalRequest, splitDetails: splitDetails || originalRequest, tasks: newTasks, completed: false, }); await this.saveTasks(); const progressTable = this.formatTaskProgressTable(requestId); return { status: "planned", requestId, totalTasks: newTasks.length, tasks: newTasks.map((t) => ({ id: t.id, title: t.title, description: t.description, })), message: `Tasks have been successfully added. Please use 'get_next_task' to retrieve the first task.\n${progressTable}`, }; }
• index.ts:41-50 (schema)

  Zod schema used for input validation of the request_planning tool parameters.
  const RequestPlanningSchema = z.object({ originalRequest: z.string(), splitDetails: z.string().optional(), tasks: z.array( z.object({ title: z.string(), description: z.string(), }) ), });
• index.ts:101-134 (registration)

  Definition of the REQUEST_PLANNING_TOOL object with name, description, and input schema.
  const REQUEST_PLANNING_TOOL: Tool = { name: "request_planning", description: "Register a new user request and plan its associated tasks. You must provide 'originalRequest' and 'tasks', and optionally 'splitDetails'.\n\n" + "This tool initiates a new workflow for handling a user's request. The workflow is as follows:\n" + "1. Use 'request_planning' to register a request and its tasks.\n" + "2. After adding tasks, you MUST use 'get_next_task' to retrieve the first task. A progress table will be displayed.\n" + "3. Use 'get_next_task' to retrieve the next uncompleted task.\n" + "4. **IMPORTANT:** After marking a task as done, the assistant MUST NOT proceed to another task without the user's approval. The user must explicitly approve the completed task using 'approve_task_completion'. A progress table will be displayed before each approval request.\n" + "5. Once a task is approved, you can proceed to 'get_next_task' again to fetch the next pending task.\n" + "6. Repeat this cycle until all tasks are done.\n" + "7. After all tasks are completed (and approved), 'get_next_task' will indicate that all tasks are done and that the request awaits approval for full completion.\n" + "8. The user must then approve the entire request's completion using 'approve_request_completion'. If the user does not approve and wants more tasks, you can again use 'request_planning' to add new tasks and continue the cycle.\n\n" + "The critical point is to always wait for user approval after completing each task and after all tasks are done, wait for request completion approval. Do not proceed automatically.", inputSchema: { type: "object", properties: { originalRequest: { type: "string" }, splitDetails: { type: "string" }, tasks: { type: "array", items: { type: "object", properties: { title: { type: "string" }, description: { type: "string" }, }, required: ["title", "description"], }, }, }, required: ["originalRequest", "tasks"], }, };
• index.ts:683-697 (registration)

  Registration of the request_planning tool (as REQUEST_PLANNING_TOOL) in the list of available tools returned by ListToolsRequest.
  server.setRequestHandler(ListToolsRequestSchema, async () => ({ tools: [ REQUEST_PLANNING_TOOL, GET_NEXT_TASK_TOOL, MARK_TASK_DONE_TOOL, APPROVE_TASK_COMPLETION_TOOL, APPROVE_REQUEST_COMPLETION_TOOL, OPEN_TASK_DETAILS_TOOL, LIST_REQUESTS_TOOL, ADD_TASKS_TO_REQUEST_TOOL, UPDATE_TASK_TOOL, DELETE_TASK_TOOL, ], }));
• index.ts:703-717 (handler)

  Entry point handler for tool calls to 'request_planning': validates input and delegates to TaskManagerServer.requestPlanning method.
  case "request_planning": { const parsed = RequestPlanningSchema.safeParse(args); if (!parsed.success) { throw new Error(`Invalid arguments: ${parsed.error}`); } const { originalRequest, tasks, splitDetails } = parsed.data; const result = await taskManagerServer.requestPlanning( originalRequest, tasks, splitDetails ); return { content: [{ type: "text", text: JSON.stringify(result, null, 2) }], }; }