Server has 25 tools with basic schemas and descriptions, but significant quality gaps. All tools have descriptions (10-50 chars, mostly generic), and input schemas are present with type definitions. However, descriptions are uniformly brief and lack context for LLM selection. No output schemas documented. No error handling guidance. No parameter validation rules or constraints. Tools lack idempotency hints, dependency documentation, or recovery guidance. Naming is verb-first and clear (list_*, create_*, get_*), which is good. But descriptions like 'Lists Intune-managed devices from your tenant' (47 chars) provide minimal context for when/why to call the tool vs alternatives. No pagination parameters on list tools despite potential for large result sets. Write operations (create_user, create_file_*) lack confirmation/dry-run patterns.

• Add pagination parameters (limit: 1-100, offset: 0+) to all 18 list_* tools. Document in descriptions: 'Returns up to 50 items per call. Use offset to fetch additional pages.'
• Expand all tool descriptions to 80-150 characters. Include WHEN to use (e.g., 'Call this first to see available Intune policies before assigning to groups'), WHAT it returns (e.g., 'Returns policy ID, name, and assignment count'), and any prerequisites.
• Document output schemas for all tools. For create_user, specify: returns {user_id, display_name, user_principal_name, created_at}. For list_* tools, specify: returns {items: [{id, name, ...}], total_count, next_offset}.
• Add error handling guidance to write operations. E.g., create_user: 'Returns 409 if UPN already exists (retryable with different UPN). Returns 403 if insufficient permissions (fatal, contact admin).'
• Add idempotency hints to create_* tools. E.g., 'Safe to retry: if user already exists with same UPN, returns existing user instead of error.' This enables agent retry logic without side effects.