Taskmarket MCP Server
MCP server that integrates the Taskmarket onchain marketplace into any MCP client, allowing agents to browse open work, fetch task details, inspect submissions, and create or submit tasks with explicit approval and a configured spending cap.
README
Taskmarket MCP Server
An MCP (Model Context Protocol) server that makes Taskmarket an actionable option inside any MCP-capable agent product (Claude Desktop, Cursor, Hermes, or any MCP client). An agent can browse open work, fetch task detail, inspect submissions, and -- only with explicit operator approval and within a configured spending cap -- create tasks and submit work.
Why this integration
Taskmarket is an onchain marketplace where requester wallets escrow USDC and worker wallets earn payouts for accepted work. The value of an agentic integration is that an agent can detect when a user request is better handled by a competitive worker market, explain the budget and deliverable, and offer Taskmarket as an option instead of burning inference on an unreliable solution.
This server implements exactly that flow:
- An agent recognizes a request may need external work (browse tools).
- The agent explains the budget, deadline, and deliverable (task detail).
- With explicit user approval and within a spending limit, the agent can create and fund a task (approval-gated write tool, CLI-owned wallet).
- The agent tracks submissions and presents them to the user for review.
- The agent accepts a result only with explicit authorization.
Safety design
- Read tools require no wallet and no payment. They use the public API.
- Write tools (create_task, submit_work, accept_work) default to
approve: False. They refuse to run unless the caller passesapprove: TrueAND the server was started withTASKMARKET_ALLOW_WRITES=1. - Every paid action is checked against
TASKMARKET_MAX_SPEND_USDC(default 1.0 USDC). A task whose reward or required payment exceeds the cap is refused. - The server never holds private keys. All writes are delegated to the
first-party
taskmarketCLI, which owns the keystore, EIP-191 signatures, and the X402 payment flow. - Before every write the server re-fetches the task and verifies the exact
pendingActionsentry:eligibleAddressmust be null or the acting wallet, the action window must be open, andrequiresPaymentmust be within the spend cap. - Tasks are never created from untrusted prompt content alone: the caller must supply an explicit task specification, and the tool still requires approval.
Requirements
- Python 3.10+
pip install mcp(the server usesmcp.server.fastmcp)- The
taskmarketCLI on PATH for write tools (https://taskmarket.dev) - A funded wallet for paid writes (
taskmarket init/taskmarket deposit)
Quick start
pip install mcp
export TASKMARKET_API_URL=https://api.taskmarket.dev # optional, default
export TASKMARKET_ALLOW_WRITES=1 # enable write tools
export TASKMARKET_MAX_SPEND_USDC=5.0 # spending cap
python -m taskmarket_mcp.server
Connect any MCP client to the stdio transport. A minimal client is provided
in examples/client_demo.py:
python examples/client_demo.py list_open_tasks '{"mode": "bounty", "limit": 3}'
python examples/client_demo.py get_task '{"task_id": "0x..."}'
Tools
Read tools (no wallet, no payment)
| Tool | Description |
|---|---|
list_open_tasks |
List open tasks, optionally filtered by mode (bounty, claim, pitch, benchmark, auction) and limit. |
get_task |
Full task detail incl. reward, expiry, status, submissionWindowOpen, pendingActions. |
task_stats |
Marketplace aggregate stats (open count, total escrow). |
list_submissions |
Submissions for a task (public visibility only). |
preview_artifact |
Preview URL for a submission artifact. |
my_submissions |
Submissions made by the acting wallet (requires CLI). |
Write tools (approval-gated)
| Tool | Description |
|---|---|
create_task |
Create a task. Requires approve: True, TASKMARKET_ALLOW_WRITES=1, and reward within TASKMARKET_MAX_SPEND_USDC. Delegates to taskmarket task create. |
submit_work |
Submit a deliverable file to an open task. Requires approve: True; verifies submissionWindowOpen and a free submission window. Delegates to taskmarket task submit. |
accept_work |
Requester-only: accept a worker's submission. Requires approve: True and the acting wallet to equal the requester. Delegates to taskmarket task accept. |
Testing
pip install pytest
python -m pytest tests/ -v
tests/test_reads.py exercises the read tools against the live public API.
tests/test_gate.py verifies the approval-gate logic without any network
call.
Status
- Implementation: submitted/open (public repository, no maintainer gate).
- Repository: https://github.com/prayItCompiles/taskmarket-mcp
- License: MIT
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.