Dooray MCP Server

Instructions

Implementation Reference

• src/tools/projects/get-task.ts:17-39 (handler)

  The handler function that implements the core logic of the 'get-task' tool. It fetches detailed task information via the projects API and returns it as formatted JSON or an error response.

  export async function getTaskHandler(args: GetTaskInput) {
    try {
      const result = await projectsApi.getTaskDetails(args.taskId, args.projectId);
  
      return {
        content: [
          {
            type: 'text',
            text: JSON.stringify(result, null, 2),
          },
        ],
      };
    } catch (error) {
      return {
        content: [
          {
            type: 'text',
            text: `Error: ${formatError(error)}`,
          },
        ],
        isError: true,
      };
    }

• src/tools/projects/get-task.ts:10-15 (schema)

  Zod schema for input validation of the get-task tool, requiring taskId and optionally projectId.

  export const getTaskSchema = z.object({
    taskId: z.string().describe('Task ID (unique identifier)'),
    projectId: z.string().optional().describe('Project ID (optional)'),
  });
  
  export type GetTaskInput = z.infer<typeof getTaskSchema>;

• src/index.ts:51-51 (registration)

  Registration of the 'get-task' tool in the central tool registry, mapping name to its handler and schema.

  'get-task': { handler: getTaskHandler, schema: getTaskSchema },

• src/index.ts:74-74 (registration)

  Inclusion of the getTaskTool in the list of available tools served by the MCP server.

  getTaskTool,

• src/tools/projects/get-task.ts:42-98 (schema)

  MCP tool specification including name, detailed description, and input schema for the get-task tool.

  export const getTaskTool = {
    name: 'get-task',
    description: `Get detailed information about a specific task.
  
  This tool retrieves complete details of a task including its full body content, attachments, workflow status, assignees, and all metadata.
  
  **URL Pattern Recognition - Two formats supported**:
  
  1. **Project-scoped URL**: "https://nhnent.dooray.com/task/PROJECT_ID/TASK_ID"
     - Extract the first numeric ID after "/task/" as projectId (optional)
     - Extract the second numeric ID as taskId (required)
     - Example: "https://nhnent.dooray.com/task/1769381697328002548/4206132384174602537"
       → {"taskId": "4206132384174602537", "projectId": "1769381697328002548"}
  
  2. **Task-only URL**: "https://nhnent.dooray.com/project/tasks/TASK_ID"
     - Extract the numeric ID after "/tasks/" as taskId
     - Example: "https://nhnent.dooray.com/project/tasks/4206132384174602537"
       → {"taskId": "4206132384174602537"}
  
  **IMPORTANT**:
  - **taskId is REQUIRED**, projectId is OPTIONAL
  - The taskId is the unique identifier (ID) from the URL, NOT the sequential task number shown in the UI
  - You can fetch task details with just the taskId without knowing the projectId
  - When a specific task URL is provided, use this tool directly instead of calling get-project-list first
  - To find taskId from task descriptions, use get-task-list to search and get task IDs
  
  Examples:
  - From URL pattern 1: {"taskId": "4206132384174602537", "projectId": "1769381697328002548"}
  - From URL pattern 2: {"taskId": "4206132384174602537"}
  - Just taskId: {"taskId": "789012345"}
  - With both IDs: {"taskId": "789012345", "projectId": "123456"}
  
  **Returns complete task information including:**
  - **Basic info**: id, number, subject, taskNumber (PROJECT-CODE/NUMBER format)
  - **Project**: project object with id and code
  - **Status**: workflowClass (registered/working/closed), workflow (id and name), closed flag, priority
  - **Dates**: createdAt, updatedAt, dueDate, dueDateFlag
  - **Content**: body with mimeType (text/x-markdown or text/html) and content
  - **Hierarchy**: parent task information (id, number, subject) if this is a subtask
  - **People**: users object with from (creator), to (assignees), cc (watchers)
  - **Organization**: milestone (id, name), tags array (id, name)
  - **Files**: files array with attachments (id, name, size)`,
    inputSchema: {
      type: 'object',
      properties: {
        taskId: {
          type: 'string',
          description: 'Task ID (unique identifier, REQUIRED)',
        },
        projectId: {
          type: 'string',
          description: 'Project ID (optional, can be omitted)',
        },
      },
      required: ['taskId'],
    },
  };