Single tool 'gemini_web_search' has a basic schema with one required parameter but critical gaps in production readiness. Description is minimal (56 chars, below 10-char baseline for context). Parameter 'query' lacks detailed constraints, format guidance, or examples. No output schema documented, LLMs cannot plan downstream operations or extract structured results. No error handling guidance, no pagination support despite web search returning variable result counts. Tool naming is acceptable (verb_noun form) but lacks specificity about what 'web search' returns or when to use it vs. alternatives. Code shows tool registration via decorator pattern but actual tool implementation (tools/gemini_web_search.py) is not provided, preventing full schema validation.

• Expand tool description to 150 - 250 characters: explain that it searches the web via Gemini API, returns ranked results with title/URL/snippet, and is best for current events or factual lookups. State any rate limits or API prerequisites.
• Add output schema documenting result structure: array of {title: string, url: string, snippet: string, source: string} with total_results count and next_cursor for pagination.
• Add 'limit' parameter (default 10, range 1 - 50) and 'offset' or 'next_cursor' for pagination. Document in description: 'Returns up to 50 results per call; use offset to fetch additional pages.'
• Enhance 'query' parameter description: 'Search query (1 - 500 characters). Supports boolean operators (AND, OR, NOT) and quoted phrases. Examples: "machine learning" OR AI, site:github.com python.'
• Add error handling: return structured errors with actionable guidance. E.g., 'API rate limit exceeded. Retry after 60 seconds.' or 'Invalid query: empty string. Provide at least one search term.'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry