mcp

The server exposes highly privileged tools to AI agents. Specifically, the 'execute_code' tool in ATP (Agent Tool Protocol) mode allows AI agents to execute arbitrary JavaScript code dynamically. This constitutes a remote code execution (RCE) vulnerability if the AI agent or the code generated by it is not fully trusted or the server environment is not adequately sandboxed and isolated. Additionally, the 'all_monday_api' tool (Dynamic API Tools) allows construction and execution of arbitrary GraphQL queries and mutations against the monday.com API, granting extensive read/write access. The server relies on a monday.com API token, which, if compromised, could lead to unauthorized access and manipulation of the user's monday.com account. Users must exercise extreme caution when deploying this server, especially when enabling ATP or Dynamic API tools.