Four tools with complete schemas, descriptions, and proper type definitions. Tool names follow verb_noun pattern (convert_document, list_conversions, get_conversion_result, search_documents). All parameters have descriptions and types. Tool annotations (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) are explicitly declared per MCP spec. Descriptions are concise (10-50 chars) but lack actionable context for LLM selection, missing WHEN to use, prerequisites, and recovery guidance. Output schemas are not documented in the source. Error handling is present but not visible in tool definitions.

• Expand tool descriptions to 100-200 chars. Include WHEN to use (e.g., 'Call list_conversions first to see pending jobs, then get_conversion_result to retrieve a specific job's output'), prerequisites, and what the tool returns.
• Document output schemas for all tools. Specify which fields are returned (e.g., conversion_id, status, created_at, result_text) so LLMs can chain tools and extract required IDs.
• Update space_id/space_name parameter descriptions to explicitly state: 'If space_id is provided, space_name is ignored. If only space_name is provided, the server resolves it to space_id via the spaces API.'
• Add pagination support to search_documents. Accept offset/limit or cursor parameters and return total_count so agents can iterate through large result sets without hitting the 5-result cap.
• Document error scenarios in tool descriptions. E.g., 'Returns error if file_path does not exist or space_id is invalid. Retry with a valid path or call list_spaces to find available space IDs.'