MCP RC Long-Running Task Prototype
Demonstrates MCP 2026-07-28 behavior for long-running tool calls, task lifecycle (get, update, cancel), and elicitation clarification during async tasks.
README
MCP RC Long-Running Task Prototype
Repository: https://github.com/StevenBucher98/mcp-rc-long-running-task-poc
Focused prototype for MCP 2026-07-28 behavior around:
- Long-running tool calls that return quickly with a task handle
- Task lifecycle methods:
tasks/get,tasks/update,tasks/cancel - Elicitation clarification while long-running work is still in progress
What Is In Scope
The server intentionally exposes only two tools:
start_report_task
- Starts an async task and immediately returns
structuredContent.task.id
personalized_greeting
- If name is not provided, returns
inputRequiredasking for a name - Continue with
requestState+inputResponses.nameto get a random greeting
Quick Start
Clone and run:
git clone https://github.com/StevenBucher98/mcp-rc-long-running-task-poc.git
cd mcp-rc-long-running-task-poc
npm install
npm run test:scenarios
Or, if you already have the repo locally:
npm install
npm run test:scenarios
The scenario runner builds and starts the server, runs the tests, prints UX-style and raw request/response output, then stops the server.
Scenario Runner
npm run test:scenarios
This single command:
- builds the server
- starts the server
- runs both scenario suites
- prints UX-style and raw JSON-RPC request/response output
- stops the server when finished
What it verifies:
- Task lifecycle basics
- create task via
tools/call(start_report_task) - update progress via
tasks/update - read status via
tasks/get - cancel via
tasks/cancel
- create task via
- Long-running behavior
- starts a long-running task (default
90000ms) - confirms tool call returns quickly (non-blocking)
- continues with other calls during execution (
personalized_greeting,tools/list) - polls until completion and prints final result
- starts a long-running task (default
Optional Flags
node scripts/test-state-and-long-running.mjs \
--long-duration-ms 90000 \
--quick-threshold-ms 4000 \
--wait-for-completion \
--greeting-name Taylor
--long-duration-ms: duration for long-running task--quick-threshold-ms: max acceptable create-call latency--wait-for-completion: wait until long-running task is complete (default behavior)--no-wait-for-completion: skip waiting at the end--greeting-name <name>: name used for greeting elicitation replay (if omitted and terminal is interactive, you will be prompted)
Endpoints
POST /mcpGET /health
Share With Others
- Repo URL: https://github.com/StevenBucher98/mcp-rc-long-running-task-poc
- To run quickly:
npm install && npm run test:scenarios
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.