This server has 18 tools with significant definition gaps. Tools 3-18 share a generic parameter set (action, vmx_path, vmx_position, verbose, output_format, command_args) that is poorly documented. Descriptions are present but generic (e.g., 'vmcli Chipset module. Valid actions: query, SetCoresPerSocket...'). No tool has documented output schemas. Parameter descriptions lack constraints, ranges, or format guidance. Error handling is not visible in tool definitions. The server acts as a thin wrapper around vmcli commands rather than providing LLM-optimized abstractions.

• Add enum constraints to the 'action' parameter for each tool. Instead of a free-form string, enumerate valid actions: e.g., for vm_power_management, action must be one of ['query', 'start', 'stop', 'pause', 'reset', 'suspend', 'unpause']. This prevents hallucinated actions.
• Document output schemas for all tools. Specify what fields are returned (e.g., vm_list returns [{vmx_path: string, vm_name: string, status: string}]). This enables agents to chain tools and extract required IDs.
• Rewrite descriptions to explain intent, not just list actions. E.g., 'Start, stop, or query the power state of a virtual machine. Use this to manage VM lifecycle.' instead of 'Power module: query, Start, Stop, Pause, Reset, Suspend, Unpause.'
• Replace 'vmx_position' with a clearer parameter name and explanation. If this controls argument order in the underlying vmcli command, abstract it away or document why the LLM must choose it.
• Add parameter constraints to 'output_format' (enum: ['json', 'yaml', 'toml']) and document what each format returns. Clarify that 'json' is recommended for structured parsing.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry