c743a10da0
Add readOnlyHint, destructiveHint, idempotentHint, and openWorldHint annotations to the tool config. Update docs to reflect that this tool now demonstrates both tool-level and content annotations. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
95 lines
3.1 KiB
TypeScript
95 lines
3.1 KiB
TypeScript
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
|
|
import { CallToolResult } from "@modelcontextprotocol/sdk/types.js";
|
|
import { z } from "zod";
|
|
import { MCP_TINY_IMAGE } from "./get-tiny-image.js";
|
|
|
|
// Tool input schema
|
|
const GetAnnotatedMessageSchema = z.object({
|
|
messageType: z
|
|
.enum(["error", "success", "debug"])
|
|
.describe("Type of message to demonstrate different annotation patterns"),
|
|
includeImage: z
|
|
.boolean()
|
|
.default(false)
|
|
.describe("Whether to include an example image"),
|
|
});
|
|
|
|
// Tool configuration
|
|
const name = "get-annotated-message";
|
|
const config = {
|
|
title: "Get Annotated Message Tool",
|
|
description:
|
|
"Demonstrates how annotations can be used to provide metadata about content.",
|
|
inputSchema: GetAnnotatedMessageSchema,
|
|
annotations: {
|
|
readOnlyHint: true, // This tool only returns data, no side effects
|
|
destructiveHint: false, // Does not delete or modify anything
|
|
idempotentHint: true, // Same input always produces same output
|
|
openWorldHint: false, // Does not interact with external systems
|
|
},
|
|
};
|
|
|
|
/**
|
|
* Registers the 'get-annotated-message' tool.
|
|
*
|
|
* The registered tool generates and sends messages with specific types, such as error,
|
|
* success, or debug, carrying associated annotations like priority level and intended
|
|
* audience.
|
|
*
|
|
* The response will have annotations and optionally contain an annotated image.
|
|
*
|
|
* @function
|
|
* @param {McpServer} server - The McpServer instance where the tool will be registered.
|
|
*/
|
|
export const registerGetAnnotatedMessageTool = (server: McpServer) => {
|
|
server.registerTool(name, config, async (args): Promise<CallToolResult> => {
|
|
const { messageType, includeImage } = GetAnnotatedMessageSchema.parse(args);
|
|
|
|
const content: CallToolResult["content"] = [];
|
|
|
|
// Main message with different priorities/audiences based on type
|
|
if (messageType === "error") {
|
|
content.push({
|
|
type: "text",
|
|
text: "Error: Operation failed",
|
|
annotations: {
|
|
priority: 1.0, // Errors are highest priority
|
|
audience: ["user", "assistant"], // Both need to know about errors
|
|
},
|
|
});
|
|
} else if (messageType === "success") {
|
|
content.push({
|
|
type: "text",
|
|
text: "Operation completed successfully",
|
|
annotations: {
|
|
priority: 0.7, // Success messages are important but not critical
|
|
audience: ["user"], // Success mainly for user consumption
|
|
},
|
|
});
|
|
} else if (messageType === "debug") {
|
|
content.push({
|
|
type: "text",
|
|
text: "Debug: Cache hit ratio 0.95, latency 150ms",
|
|
annotations: {
|
|
priority: 0.3, // Debug info is low priority
|
|
audience: ["assistant"], // Technical details for assistant
|
|
},
|
|
});
|
|
}
|
|
|
|
// Optional image with its own annotations
|
|
if (includeImage) {
|
|
content.push({
|
|
type: "image",
|
|
data: MCP_TINY_IMAGE,
|
|
mimeType: "image/png",
|
|
annotations: {
|
|
priority: 0.5,
|
|
audience: ["user"], // Images primarily for user visualization
|
|
},
|
|
});
|
|
}
|
|
|
|
return { content };
|
|
});
|
|
};
|