Skip to main content
Glama
langd0n-labs

Agenda

by langd0n-labs

MCP Servers: The Agentic API Layer

Workshop repo for Agentic 101, Washington DC, 2026-09-15.

We build an MCP server over the conference's own agenda, then use it from Copilot in agent mode, then work out when you should not have bothered.

Start here

Open this repo in a GitHub Codespace. Dependencies install themselves. Then:

  1. Open the Copilot Chat panel and put it in Agent mode.

  2. VS Code asks you to trust the MCP server this workspace defines. Say yes. The server does not start until you do.

  3. Ask: what time is the MCP talk?

If nothing happens, run MCP: List Servers from the command palette and check the server's output log.

Copilot calls list_sessions on the server in server.py, which is already wired up in .vscode/mcp.json. That is an MCP server. There isn't more to it.

Related MCP server: Sessionize

What's here

File

What it is

server.py

the MCP server, ~15 lines of it

agenda.json

today's schedule

agenda_direct.py

the same lookup with no protocol, for comparison

.vscode/mcp.json

how the editor finds the server

steps/

finished states, if you fall behind

Checkpoints

If your copy stops working, or you want to skip ahead, copy a finished state over server.py and reload the server:

cp steps/1-sessions_after.py    server.py
cp steps/2-with-tool-error.py   server.py
cp steps/3-with-description.py  server.py

See steps/README.md for what each one is.

Note on the SDK

This uses mcp 2.x, where the server class is MCPServer:

from mcp.server.mcpserver import MCPServer

Most examples you will find online are 1.x and use FastMCP. They will not run against this. The rename is recent :)

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    C
    maintenance
    MCP server for accessing Sessionize event data. Query speakers, sessions, and schedules from any Sessionize-powered conference using natural language.
    6
    4
    Apache 2.0
  • F
    license
    Not graded
    quality
    Not graded
    maintenance
    Enables delegation of immediate conversational messages and long-running background tasks to OpenClaw agents. It allows users to spawn research or monitoring tasks, check execution status, and manage active agent sessions within MCP-compatible clients.
    -
  • A
    license
    A
    quality
    D
    maintenance
    MCP server for the Indico meeting agenda system, enabling AI agents to search event categories, browse agendas, and extract contribution and session details. Works with multiple Indico instances simultaneously.
    19
    3
    MIT