Single tool with clear naming (search_seerah) and comprehensive description (287 chars, within 10-1024 baseline). Input schema properly defined with typed parameters and descriptions. However, output schema is undocumented, callers cannot predict the structure of returned passages. No error handling guidance for missing KB or invalid queries. Parameter descriptions are adequate but lack format constraints (e.g., max query length, num_results bounds enforcement). Tool is read-only and well-scoped, but lacks production-grade robustness patterns.

• Document output schema: return structured JSON with fields like {passages: [{relevance: float, source: string, content: string, truncated: bool}], total_results: int}. This lets LLMs parse results reliably.
• Add error handling: return {error: string, recovery: string} for missing KB or invalid input. E.g., 'Knowledge base not found. Run setup.sh to initialize.' This guides LLM recovery.
• Constrain num_results: update description to 'Integer between 1 and 12 (default 6). Controls number of passages returned.'
• Constrain query: add to description 'String, max 500 characters. The question or topic to search for in the Seerah knowledge base.'
• Clarify scope: expand description to explain 'Seerah' (biography of Prophet Muhammad) and when to use this tool (Islamic history, Prophet's life, battles, companions) vs. general knowledge.