Working Machines MCP Server

io.github.klei30Published by GitHub account klei30v1.0.4Updated Oct 24, 2026

Connect to 1,400+ apps and 15,000+ actions through one OAuth-protected MCP server.

VerifiedStreamable HTTPWeb executableProductivity & WorkflowCommunication & CollaborationBusiness & Commerce

Overview

AI-generated overview

A hosted MCP server that lets an assistant discover and run actions across 1,400+ connected apps, such as Slack, GitHub, Jira, Salesforce and Snowflake, through

What you can do
Discover available apps and integrations, and check which accounts are connected and what permissions they hold. Search actions by intent and retrieve validated schemas and input parameters for a chosen action. Execute actions against connected providers behind the service's execution boundary and return a structured, auditable result. Grouped by target: apps and connections (list_apps, list_connections); actions (search_actions, get_action_guide, execute_action).
When to use it
Useful when an assistant should act across many SaaS tools, such as creating issues, sending messages or querying data, without setting up a separate integration for each provider. Best for teams that want one OAuth connection and policy-controlled execution instead of many local API keys.
Requirements
A remote MCP endpoint at app.workingmachines.dev, reached over streamable HTTP. A Working Machines account with OAuth, supplied as an Authorization header in the form 'Bearer '. Accounts must be connected to the service with provider-native scopes before actions can run. Network access is required.
Before you install
The Authorization bearer token is a secret and grants access to every connected account, so keep it out of shared logs and prompts. Actions can write, send or change data in third-party apps such as Slack, GitHub, Jira, Stripe and Salesforce, and may incur costs in services like AWS or Stripe. Execution policies can require approval; review them before enabling unattended runs. Action records are kept by the hosted service.

Installation

In SourceWeft

  1. Open Working Machines MCP Server in the dashboard and add it to a workspace.
  2. Connect the required credentials. They are stored privately and never shown on public pages.
  3. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.klei30%2Fworking-machines/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config. The client completes OAuth sign-in on first connect.

{
  "mcpServers": {
    "working-machines": {
      "type": "http",
      "url": "https://app.workingmachines.dev/mcp"
    }
  }
}

README

Working Machines MCP Server for Cursor

Official Cursor plugin repository for Working Machines.

Connect Cursor to 1,400+ apps and 15,000+ machine-ready tools remotely through one secure, hosted Model Context Protocol (MCP) server.

[Working Machines Logo]

Remote Access to 1,400+ Apps & 15,000+ Tools

Working Machines provides the application layer for AI agents. Through a single remote MCP connection (https://app.workingmachines.dev/mcp), your agent gets instant access to 1,400+ integrations and 15,000+ structured actions across:

  • Team Collaboration: Slack, Gmail, Google Workspace, Microsoft Teams, Notion
  • Development & Cloud: GitHub, GitLab, Sentry, Jira, AWS, Vercel, Datadog
  • Projects & Productivity: Linear, Asana, Trello, ClickUp, Monday.com
  • Commerce & CRM: Salesforce, HubSpot, Stripe, Shopify, Zendesk
  • Data & Analytics: Snowflake, BigQuery, PostgreSQL, Airtable, Pinecone

Safety & Security Architecture

Working Machines is engineered from the ground up for safe, enterprise-grade AI execution:

🛡️ 1. Your Credentials Are Not Context

  • Zero Local Secret Storage: API keys, OAuth tokens, and secret credentials stay behind the Working Machines execution boundary.
  • Context Protection: Credentials never enter your Cursor chat context, prompt buffers, or local workspace files.

🔒 2. Scoped Capabilities & Policy Enforcement

  • Intent-Based Access: Agents discover only the minimal actions required for the task at hand rather than loading full provider schemas.
  • Provider-Native Scopes: Connect accounts once with provider-bound permissions (e.g. create issues only).
  • Execution Policies: Allow, reject, or require approval for actions before any call leaves the execution boundary.

📋 3. Verifiable & Auditable Run History

  • Structured Proof: Every action execution produces a structured, redacted, and auditable record.
  • Execution Visibility: Tracks who acted, which connection was used, what changed, timing, and policy decisions for complete transparency.

Repository Structure

This repository follows the official Cursor plugin marketplace format:

text
working-machines/├── .cursor-plugin/│   └── marketplace.json├── plugins/│   └── working-machines/│       ├── .cursor-plugin/│       │   └── plugin.json│       ├── assets/│       │   └── logo.png│       ├── skills/│       │   └── working-machines/│       │       └── SKILL.md│       ├── mcp.json│       └── README.md├── scripts/│   └── validate-template.mjs├── README.md├── CHANGELOG.md├── LICENSE├── NOTICE└── .gitignore

Agent Workflow

Cursor agents follow the Working Machines discovery and execution sequence:

text
list_apps      ↓list_connections      ↓search_actions      ↓get_action_guide      ↓execute_action      ↓verify result
  1. list_apps: Discover available applications and tool integrations.
  2. list_connections: Verify active account authorizations and permissions.
  3. search_actions: Find machine actions by intent.
  4. get_action_guide: Retrieve validated schemas and input parameters.
  5. execute_action: Execute the action securely behind the boundary.
  6. Verify Result: Audit the structured execution response.

Validation

Verify repository manifest and component compliance:

bash
node scripts/validate-template.mjs

Documentation & Links

License & Notice

  • Open-source plugin wrapper licensed under the MIT License.
  • See the NOTICE file regarding proprietary hosted infrastructure details.

Source: README.md at commit 41c220e

Tools

0
This server asks you to sign in, so SourceWeft could not read its tools without an account and has not indexed them. What a connected account sees can depend on that account and its permissions.

Version history

1
  1. v1.0.4LatestSep 16, 2026

Score

75% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

    Declare a license in the repository or the registry entry.