mcp -multi-tools

The MCP Server Multi Tools provides a unified bridge that connects AI agents to DevOps and collaboration platforms such as Azure DevOps, Slack, GitHub, and other agent-backed tools through a single, standardized MCP interface. It supports a wide range of capabilities, including work item management (create, read, update, search, comment), sprint management, WIQL queries, enrichment of items with contextual data from GitHub, Slack, and Sentry, as well as fetching repository content. In addition, it introduces an Advanced Agent System featuring Agents-in-Agents, allowing your AI agents to spawn and orchestrate nested agents inside Docker containers with web browser capabilities and iterative workflows, enabling sophisticated multi-step automation. To use it, run the server and point your MCP client to the server’s endpoint, then configure the client’s mcpServers entry with the server name and executable path or container command. The client can then send tool calls and receive structured results that orchestrate the connected services through the shared MCP protocol. For day-to-day use, you can leverage the go-based binary (or the container image) to start the service, and utilize the provided environment-variable-driven configuration to supply credentials and access tokens for Azure DevOps, Slack, and GitHub, enabling seamless integration across your devops and collaboration toolchain.

Prerequisites:

• Docker (for containerized deployment) or Go tooling if building locally
• Access tokens/credentials for Azure DevOps, Slack, and GitHub
• Basic shell environment (bash/zsh) on your system

Option A — Install via Docker (recommended for quick start):

1. Ensure Docker is running on your machine.
2. Pull and run the container image:

# Pull and run the MCP Server Multi Tools container
docker run -d --name mcp-multi-tools \
  -p 8080:8080 \
  -e AZURE_DEVOPS_TOKEN="your-azure-devops-token" \
  -e SLACK_TOKEN="your-slack-token" \
  -e GITHUB_TOKEN="your-github-token" \
  theapemachine/mcp-server-multi-tools:latest

3. The server will expose its MCP interface on port 8080 by default (adjust as needed).

Option B — Build and run locally (Go-based binary):

1. Prerequisites: Go installed, repository cloned.

# Clone the repository
git clone https://github.com/theapemachine/mcp-server-multi-tools.git
cd mcp-server-multi-tools

# Build the server binary
go build -o mcp-server-multi-tools .

# Run the binary (example with env vars)
AZURE_DEVOPS_TOKEN=your-token SLACK_TOKEN=your-token GITHUB_TOKEN=your-token \
./mcp-server-multi-tools

2. If you prefer, use the provided start.sh wrapper to initialize environment variables before launching the binary.

Prerequisites recap:

• Go (if building locally) or Docker (for containerized deployment)
• Credentials for Azure DevOps, Slack, and GitHub
• Optional: Docker Desktop for Agents-in-Agents feature

Notes and tips:

• The server is configured primarily via environment variables for flexibility; keep credentials secure and use secret management where possible.
• If you’re using the Agents-in-Agents feature, ensure Docker Desktop is running to host nested agents with Debian-based containers and embedded web browser capabilities.
• When configuring MCP clients, reference the server via its entry name (e.g., multi-tools) and provide the command or container invocation as shown in the mcp_config example.
• If you encounter connectivity issues with tokens, verify scope permissions and token validity for Azure DevOps, Slack, and GitHub.
• Logs can help diagnose issues; increase MCP_LOG_LEVEL to debug for deeper visibility during setup.
• For CI environments, prefer container deployment to ensure consistent runtime environments across hosts.