Two tools with well-structured schemas and clear descriptions. web_search has comprehensive parameter constraints (enums, min/max, patterns); web_fetch is simpler but properly defined. Both descriptions are concise and actionable (60-80 chars). Schemas use proper JSON Schema with types, descriptions, and validation rules. No critical gaps, but missing output schema documentation and error handling guidance. Tool names follow verb_noun convention clearly.

• Document output schemas for both tools. Example: web_search returns {results: [{title, url, snippet, domain}], total_count}; web_fetch returns {text, url, truncated}.
• Add error recovery hints to descriptions: 'If no results, try removing filters or broadening date range' and 'If fetch fails, verify URL is publicly accessible.'
• Clarify the relationship between tools: 'Use web_search to find URLs, then web_fetch to extract full text from promising results.'
• Add guidance on web_fetch pagination: 'Use offset to retrieve subsequent chunks; set max_chars lower (5000-10000) for faster responses.'
• Enhance web_search description to explain handling of untrusted data: 'Verify critical facts with multiple sources; do not assume snippets are accurate.'