Four tools with complete input schemas and descriptions. Naming follows verb_noun pattern (tavily-search, tavily-extract, tavily-crawl, tavily-map). Descriptions are substantive (150-250 chars), exceeding the 10-char minimum but below the 50-200 char LLM-optimized range. All parameters have type definitions and descriptions. However, output schemas are not documented, LLMs cannot predict response structure. Error handling is minimal; no recovery guidance or actionable error messages visible in code. Deidentification parameters add complexity without clear necessity for most use cases. No tool annotations (readOnlyHint, destructiveHint) despite all tools being read-only.

• Document output schemas for all four tools. Example for tavily-search: {type: 'object', properties: {query: string, results: [{title: string, url: string, content: string, score: number, published_date?: string}], answer?: string, images?: string[]}, required: ['query', 'results']}
• Add tool annotations: readOnlyHint=true and idempotentHint=true to all four tools. This signals to agents that retries are safe and no state is modified.
• Implement error handling with actionable messages. Example: 'API rate limit exceeded. Retry after 60 seconds.' or 'Invalid query: must be 1-500 characters. Got: "' + query.length + ' chars"'
• Move Skyflow deidentification to server-side configuration (environment variables) or a separate optional tool. Remove enable_deidentification and deidentify_entities from tavily-search, tavily-extract, tavily-crawl, tavily-map to reduce parameter complexity.
• Add conditional schema validation: if topic='news', require time_range or days; if topic='general', reject days parameter. Document this in parameter descriptions.

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

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