Backlog MCP Server

Instructions

Implementation Reference

• src/tools/getDocuments.ts:36-38 (handler)

  The handler function that executes the tool logic by calling the Backlog API's getDocuments method with provided projectIds and offset.

  handler: async ({ projectIds, offset }) => {
    return backlog.getDocuments({ projectId: projectIds, offset });
  },

• src/tools/getDocuments.ts:7-18 (schema)

  Input schema definition for the get_documents tool, including projectIds (array of project IDs) and optional offset for pagination.

  const getDocumentsSchema = buildToolSchema((t) => ({
    projectIds: z
      .array(z.number())
      .describe(t('TOOL_GET_DOCUMENTS_PROJECT_ID_LIST', 'Project ID List')),
    offset: z
      .number()
      .optional()
      .default(0)
      .describe(
        t('TOOL_GET_DOCUMENTS_OFFSET', 'Offset for pagination (default is 0)')
      ),
  }));

• src/tools/tools.ts:151-155 (registration)

  Registration of the getDocumentsTool within the 'document' toolset group in the allTools function.

  tools: [
    getDocumentsTool(backlog, helper),
    getDocumentTreeTool(backlog, helper),
    getDocumentTool(backlog, helper),
  ],

• src/tools/getDocuments.ts:29-35 (schema)

  Tool metadata including name 'get_documents', description, input/output schemas, and important fields.

  description: t(
    'TOOL_GET_DOCUMENTS_DESCRIPTION',
    'Gets a list of documents in a project.'
  ),
  schema: z.object(getDocumentsSchema(t)),
  outputSchema: DocumentItemSchema,
  importantFields: ['id', 'projectId', 'title', 'plain'],