Single tool 'synthesize_page' has critical definition gaps. Tool description is present (67 chars) but lacks actionable context on WHEN to use it or WHAT it returns. Input schema visible in code but parameters lack individual descriptions, 'entities' and 'ranking_weights' have no guidance on format, constraints, or expected structure. No output schema documented. No error handling guidance. Tool does not follow verb_noun naming (should be 'synthesize_comparison_page' or similar). No parameter validation rules, no enum constraints, no range limits. Fails 6 of 7 critical checks in DEFINITION QUALITY rubric.

• Rename tool to 'rank_products_by_composite_score' or 'create_product_comparison_matrix' to signal action verb and intent clearly.
• Add parameter descriptions: 'entities' → 'Array of product objects, each with sku (string), name (string), price (float), rating (0-5), specs (object), pros (array), cons (array). If omitted, uses default robot vacuum dataset.' 'ranking_weights' → 'Object with keys price_importance, rating_importance, feature_richness (each 0-1, sum should equal 1.0). Defaults to {price: 0.35, rating: 0.45, feature: 0.20}.'
• Document output schema: 'Returns object with status (string), winner_name (string), winner_score (float 0-10), ranked_entities (array of input entities plus composite_score field), comparison_table_markdown (string), html_embed_snippet (string), total_evaluated (integer).'
• Expand tool description to 100-150 chars: 'Ranks and compares products by composite score (weighted price, rating, feature count). Returns ranked list, markdown comparison table, and HTML recommendation card. Use when you have 2+ products to evaluate and need a structured comparison.'
• Add error handling: validate entities array is non-empty, each entity has required fields (sku, name, price, rating), ranking_weights values are 0-1, and sum is non-zero. Return clear error messages like 'Invalid entities: missing required field "price" in entity at index 2' instead of silent defaults.