Scoring was not performed

• Add non-empty descriptions to all 11 tools currently lacking them. Aim for 50 - 200 characters explaining what the tool does, when to use it, and what it returns.
• For each parameter, add a description that specifies the expected format, valid values or range, and any constraints (e.g., 'A command string: one of [start, stop, pause]. Required.' or 'A file name (string, max 255 chars)')
• Document output schemas for all tools. Specify the structure of successful responses and error responses. Use JSON Schema format consistent with input schemas.
• For WRITE-risk tools (long_running_task, generate_business_card, pay_a_bill), add explicit descriptions of side effects: 'This tool will execute a command on the server. Irreversible.' or 'This generates a business card and may store it in the database.'
• Rename generic tools to follow verb_noun_resource pattern: 'trace_tool' → 'get_trace_logs', 'admin_tool' → 'call_admin_function', 'hello_world' → 'greet_user'
• Add enum constraints or format patterns for parameters that accept only a limited set of values. For 'command' parameter in long_running_task, specify the exact list of valid commands.
• Include error handling guidance: what does the tool return on failure? What should the agent do next? Example: 'If the resource is not found, returns error code 404. Try search_resources() with a partial name.'
• Consider adding a pagination pattern to tools that return large result sets (if any exist). Specify limit, offset/page, and a total count in responses.