This server exposes a single tool 'execute' that runs arbitrary shell commands. The tool has a minimal description ('Execute planning command') and a basic input schema with only a 'command' string parameter. No output schema is documented. The tool lacks parameter descriptions, error guidance, and security controls. This is a destructive, high-risk tool with no safeguards, confirmation steps, or audit logging. The implementation directly calls child_process.exec() without sanitization, making it vulnerable to command injection. No MCP protocol features are implemented beyond basic HTTP.

• Rename tool to 'run_shell_command' or 'execute_bash_command' to clarify intent and follow verb_noun convention.
• Expand description to 100-200 chars: 'Execute a shell command on the server. Returns stdout and stderr. WARNING: This tool can execute any command, use only for safe, pre-approved operations. Supports bash, sh, and standard Unix utilities.'
• Add parameter description for 'command': 'The shell command to execute (string). Must be a valid bash/sh command. Examples: ls -la, echo hello, grep pattern file. Avoid destructive commands (rm, kill, dd) unless explicitly required.'
• Document output schema in tool description or as a separate schema field: 'Returns {stdout: string (command output), stderr: string (error output), error?: string (execution error if command failed)}'.
• Implement a whitelist of allowed commands or a dry-run mode. At minimum, add a 'confirm' boolean parameter that requires explicit confirmation before executing destructive operations (commands containing rm, kill, dd, etc.).