Server has 5 tools with complete schemas and descriptions. Naming follows verb_noun convention (search_, run_, list_, get_). Descriptions are present (avg 150 chars) and explain WHAT the tool does. However, parameter descriptions lack actionable constraints (no ranges, patterns, or format guidance). Output schemas are not documented, LLMs cannot predict response structure. Error handling is minimal; no recovery guidance. Tool composition is sound (each does one thing), but missing idempotency hints and chaining IDs in responses.

• Document output schemas for all tools. Example for toad_search_documents: 'Returns array of {id, title, content, score, url}. Score is 0-1 relevance. Limit results to 50 max.'
• Add format and constraint hints to parameter descriptions. E.g., 'query: Natural language search (1-500 chars, no special operators)' and 'limit: Results per page (1-50, default 10, higher values slow response)'.
• Add error handling guidance to tool descriptions. E.g., 'If Semantic Search API is unavailable, returns error 503. Retry after 5 seconds or call toad_system_status to check health.'
• For toad_run_eval, add idempotency hint: 'Running the same suite twice with identical parameters returns the same results. Safe to retry on transient failures.'
• Ensure response objects include all IDs and references needed by downstream tools. E.g., toad_list_prompts should return {name, version, description, ...} so agents can immediately call toad_get_prompt with the name.