Three tools with complete input schemas and descriptions, but descriptions are generic and lack LLM-optimized guidance. All parameters have types and descriptions, meeting baseline requirements. However, output schemas are not documented, error handling is absent, and parameter constraints (enums, ranges) are minimal. Tool naming follows verb_noun convention (search_, get_, get_competitors) but lacks clarity on when to use each. No security considerations documented despite handling business data.

• Document output schema for each tool. Example: search_businesses returns {results: [{id, name, phone, website, suburb, state, country, google_rating, quality_score}], total_count: int}. This enables LLMs to chain get_business(id) calls.
• Convert 'directory' and 'country' parameters to enums. Declare all 21 valid directories and 8 country codes in the schema. Remove example lists from descriptions.
• Add numeric constraints to 'limit' and 'min_rating'. E.g., limit: {type: integer, minimum: 1, maximum: 50, default: 10}; min_rating: {type: number, minimum: 0, maximum: 5}.
• Expand tool descriptions with WHEN-to-use guidance. E.g., 'Use search_businesses to find providers by name/type/location. Use get_competitors to analyze rivals in the same category within a suburb.'
• Add error handling guidance. E.g., 'If no results found, try broadening the suburb or removing filters. If directory is invalid, call search_businesses without directory filter to discover available categories.'