Server has 7 read-only tools with consistent naming (show_*) and reasonable descriptions (100-150 chars). All tools have input schemas with type definitions and parameter descriptions. However, output schemas are not documented, LLMs cannot predict response structure. Parameter descriptions lack format constraints (e.g., interface naming quirks are mentioned but not formalized). No error handling guidance or recovery paths. Missing tool annotations (readOnlyHint present conceptually but not declared). Baseline: 194 chars avg description, 72 chars avg param annotation, this server meets description baseline but param annotations are sparse.

• Document output schema for each tool. Example: show_interface_brief returns {switch: string, response: {jsonrpc, result: {body: {interface_name, state, speed, ...}}}}. This enables LLMs to extract and chain data.
• Add readOnlyHint: true to all tool schemas. FastMCP supports tool annotations via the schema, declare them explicitly so clients can render safety indicators.
• Formalize interface parameter constraints as a regex pattern in the schema (e.g., pattern: '^(fc\d+/\d+(/\d+)?|port-channel\d+|po\d+)$') rather than relying on description text.
• Provide actionable error messages. When zone name validation fails, return 'Invalid zone name: must start with letter, contain only [A-Za-z0-9_$^-], max 64 chars. Got: "xyz-123"' instead of a generic ValueError.
• Add a discovery tool (e.g., list_zones, list_interfaces) to help LLMs explore available resources before querying specific ones. This reduces failed lookups and improves planning.