Server provides 10 read-only tools with complete JSON schemas and descriptions. All tools follow verb_noun naming (buscar_*, obtener_*) and have parameter descriptions. However, descriptions are minimal (avg 80 chars, below 194-char baseline), lack WHEN/WHY context, and omit error recovery guidance. No output schemas documented. Parameters lack format constraints (e.g., date format YYYY-MM-DD stated in description but not in schema). No tool annotations (readOnlyHint present in metadata but not in schema). Error handling is basic, no guidance on retryability or next steps.

• Expand tool descriptions to 100 - 200 chars. Include WHEN to use (e.g., 'Use this when you have a case number (rol) and need full details'), WHAT it returns (e.g., 'Returns case metadata, parties, and timeline'), and error recovery (e.g., 'If rol is invalid, try buscar_causas_por_nombre with party name').
• Document output schemas for all tools. Specify whether search tools return arrays or paginated objects, what fields are included, and which IDs/references are available for chaining (e.g., does buscar_causas_por_nombre return rol so obtener_detalle_causa can be called next?).
• Add 'format: date' to date parameters (fecha_desde, fecha_hasta) in JSON Schema. Remove format examples from descriptions and rely on schema constraints.
• Add pagination to search tools: accept 'limit' (default 20, max 100) and 'offset' (default 0) parameters. Return 'total_count' and 'next_offset' in response so LLMs can iterate without guessing.
• Expose readOnlyHint in tool input schemas or via a tool_annotations capability. This signals to LLMs that all tools are safe to retry and have no side effects.