Sent Headers in MCP sse server

[04cfb1ed]

Validations

• I believe this is a way to improve. I'll try to join the Continue Discord for questions
• I'm not able to find an open issue that requests the same enhancement

Problem

Hello, I'm trying to use an MCP sse server protected by authentication, but I found that the headers are not sent in the request

I did the change in a local environment, and I want to share with you, and if it's useful could be added in the code

Solution

Headers in interfaces
https://github.com/continuedev/continue/blob/main/core/index.d.ts#L1105-L1124

export interface StdioOptions {
  type: "stdio";
  command: string;
  args: string[];
  env?: Record<string, string>;
  headers?: HeadersInit;
}

export interface WebSocketOptions {
  type: "websocket";
  url: string;
  headers?: HeadersInit;
}

export interface SSEOptions {
  type: "sse";
  url: string;
  headers?: HeadersInit;
}

Send headers to mcp library
https://github.com/continuedev/continue/blob/main/core/context/mcp/index.ts#L170-L171

case "sse":
        console.log("HEADERS");
        console.log(options.transport.headers);
        return new SSEClientTransport(new URL(options.transport.url), {
          eventSourceInit: {
            fetch: (input, init) =>
              fetch(input, {
                ...init,
                headers: {
                  ...init?.headers,
                  ...(options.transport.headers as Record<string, string> | undefined),
                }
              }),
          },
          requestInit: { headers: options.transport.headers }
        });

These changes allow using a config like this (config.json)

"experimental": {
        "modelContextProtocolServers": [
            {
                "name": "hola",
                "transport": {
                    "type": "sse",
                    "url": "https://protected.example.org/sse",
                    "headers": {
                        "Authorization": "Bearer supertoken"
                    }
                }
            },
        ]
    }

[Patrick-Erichsen]

This would definitely be a welcome PR! Mind opening one? Your logic looks sound to me.

[04cfb1ed]

I opened the PR #5493 with the changes proposed