mcp-agent-server
by mbarki-abd
Overview
The MCP Agent Server orchestrates a multi-agent AI platform, enabling the creation, management, communication, and real-time monitoring of AI agents, their workspaces, projects, tasks, and secure credentials.
Installation
docker compose up -dEnvironment Variables
- MCP_API_KEY
- ENCRYPTION_KEY
- DB_NAME
- DB_USER
- DB_PASSWORD
- ANTHROPIC_API_KEY
- GOOGLE_API_KEY
- OPENAI_API_KEY
- MISTRAL_API_KEY
- OLLAMA_HOST
Security Notes
CRITICAL VULNERABILITIES IDENTIFIED: 1. **Information Leak on Public Endpoints**: The `/ws/dashboard`, `/api/dashboard/state`, `/api/presence/ws`, and `/api/presence/sse` endpoints are explicitly marked as public routes in `auth.ts`. These endpoints expose sensitive system information including server details, a list of all agents, their statuses, active tasks, system metrics, projects, and session metadata, all without any authentication. This allows unauthenticated users to gain comprehensive visibility into the operational state of the entire multi-agent platform. 2. **Secret Master API Key Exposure**: The `/key` endpoint (path configured by `SECRET_KEY_PATH`) returns the master API key. While the code includes a warning to 'keep this path secret', relying on obscurity for access to a critical secret is a significant security risk. If this path is discovered (e.g., through scanning, misconfiguration, or leaks), the master API key is exposed. 3. **Root Privileges Required**: Core functionalities such as agent provisioning (creating Unix users and their home directories), setting file ownership (`chown`), and running commands as other users (`su`) require the server to operate with root privileges or highly permissive sudoer configurations. This introduces a substantial attack surface, as a compromise of the server could lead to full system compromise. RECOMMENDATIONS: Implement robust authentication/authorization for all dashboard and presence endpoints. Remove or severely restrict the `/key` endpoint, ideally replacing it with a secure token issuance/rotation mechanism for master access.
Similar Servers
klavis
Develop and deploy AI agents that interact with a wide array of web services (e.g., Gmail, YouTube, LinkedIn, Supabase, Salesforce, Kubernetes) through a standardized Model Context Protocol (MCP), often orchestrated by an intelligent routing layer like Strata.
ultimate_mcp_server
The Ultimate MCP Server acts as a comprehensive AI agent operating system, providing advanced AI agents access to a rich ecosystem of tools, cognitive systems, and specialized services via the Model Context Protocol for cognitive augmentation, tool use, and intelligent orchestration.
ironmanus-mcp
Orchestrates AI workflows with an 8-phase control flow and specialized tools, serving as a Model Context Protocol (MCP) server.
mcpc
Build and compose agentic Model Context Protocol (MCP) servers and tools, enabling AI assistants to discover, integrate, and orchestrate other MCP servers for complex tasks.