Three minimal tools with bare-bones definitions. All have descriptions (10-20 chars, below the 20-char floor), no parameter descriptions, and schemas lacking detail. 'echo' has one parameter but no description for it. 'ping' and 'whoami' have empty property objects. No output schemas documented. No error handling guidance. This is a demo/audit harness, not a production tool suite, it deliberately uses trivial, side-effect-free tools to test MCP server authorization boundaries, not to provide useful agent capabilities.

• Expand tool descriptions to 50-150 chars, explaining WHAT each does, WHEN to call it, and WHAT it returns. E.g., 'echo: Echoes back the supplied text unchanged. Use to test connectivity or verify text transmission. Returns the input text as-is.'
• Add parameter descriptions. For 'echo', document: 'text (string): The text to echo back. Any length. Returns unchanged.'
• Document output schemas. E.g., 'echo returns: {type: string, description: "The echoed text"}'. 'ping' and 'whoami' return strings, make this explicit.
• Add error handling guidance. E.g., 'If echo receives null or undefined text, returns empty string. If whoami fails, returns error message with recovery hint.'
• Consider renaming tools to action-verb form if used in production: 'return_text' (echo), 'test_server' (ping), 'get_server_identity' (whoami). Current names are too generic for LLM intent inference.