Deep Research has 4 tools with minimal schema documentation and weak descriptions. tavily_search and think_tool have single-line descriptions (20-50 chars, below the 50-200 char baseline for LLM-optimized descriptions). ConductResearch and ResearchComplete lack parameter type definitions in visible code. No output schemas documented. No error handling guidance. Tool names lack clear verb-noun patterns (think_tool, ResearchComplete are vague). No evidence of input validation, permission gates, or audit trails. This is a D-tier server: major gaps in schema completeness, description quality, and error handling.

• Expand tavily_search description to 100-150 chars: 'Search the web for current information on a topic. Returns ranked results with title, URL, and snippet. Use when you need recent data or external sources not in training data. Fails if query is empty or API rate limit exceeded.'
• Rename think_tool to reflect_on_progress or plan_next_steps. Add description: 'Pause to reflect on research progress, identify gaps, and plan next steps. Use between search rounds to synthesize findings and decide what to research next.'
• Document output schemas for all tools. E.g., tavily_search returns [{title, url, snippet, source}]; ConductResearch returns {findings, sources, confidence_score}.
• Add error handling to descriptions. E.g., 'If no results found, try broader keywords or different search terms. If rate limited, wait 60s before retrying.'
• Rename ResearchComplete to mark_research_complete or finalize_research. Add description: 'Signal that research is complete and ready for final report generation. Call this after all research questions are answered.'