This server has three tools with READ_ONLY operations. All three have basic descriptions and input schemas, but fall short of production-grade quality. Tool names follow verb conventions (search_, get_, tavily-search), but descriptions are generic and lack depth. No parameter descriptions are visible in the schema definitions provided. Output schemas are not documented. Error handling is not evident from the code. The tools are simple wrappers around external APIs (ArXiv, Tavily), so the implementation complexity is low, but the interface quality is below the baseline for A-grade tools.

• Add parameter descriptions to all tools. For 'query', specify: 'Search query string (1-500 characters). Use keywords or natural language phrases. For ArXiv: try author names, paper titles, or technical terms (e.g., "transformer attention mechanism").'
• Document output schemas for each tool. E.g., 'search_arxiv returns: {results: [{paper_id, title, authors, published_date, abstract}], total_count, next_page_token}'
• Expand tool descriptions to 150-200 chars. For search_arxiv: 'Search the ArXiv preprint repository for academic papers by title, author, or keywords. Returns paper summaries with IDs, abstracts, and publication dates. Use this to find relevant research; then call get_arxiv_paper to retrieve full text.'
• Add pagination guidance to search tools. E.g., 'Limit results to max 20 per request to reduce token usage. Use page_number or next_token to fetch additional results.'
• Add error handling examples to descriptions. E.g., 'If no papers match, try a simpler query with fewer keywords. If paper_id is invalid, call search_arxiv first to find the correct ID.'
• Rename 'tavily-search' to 'search_web' for consistency with other tool names and to clarify it is a web search capability, not Tavily-specific.

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

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