Server provides 9 tools with complete JSON schemas and basic descriptions. All tools have input schemas with typed parameters and defaults. However, descriptions are generic (avg ~60 chars, well below the 194-char baseline for A+ tools) and lack context about WHEN to use each tool or what distinguishes them. Parameter descriptions are minimal. No output schemas documented. Error handling returns generic error dicts without recovery guidance. Missing tool annotations (readOnlyHint, destructiveHint) despite clear risk levels. No idempotency guarantees stated. Composition is sound (single responsibility per tool), but descriptions need expansion to meet LLM-optimization standards.

• Expand tool descriptions to 100-200 chars. Include WHEN to use (e.g., 'Use search_read when you need both IDs and full records in one call; use search alone if you only need IDs to minimize data transfer'). Reference pattern:tool-description.
• Add tool annotations to all tools: readOnlyHint=true for search/read/search_read/search_count/fields_get/default_get; destructiveHint=true for unlink; idempotentHint=true for read/search/fields_get/default_get. Reference pattern:tool-annotations.
• Document output schemas in tool descriptions or as a separate schema field. E.g., 'Returns {records: [{id, name, ...}], count: int, model: string}'. Reference pattern:response-shaper.
• Enhance error responses with categorization and recovery hints. E.g., 'PermissionError: User lacks write:res.partner permission. Contact admin or use a different model.' Reference pattern:recovery-guide.
• Add a confirm_unlink tool or dry_run parameter to unlink/write/create to prevent accidental data loss. Reference pattern:confirmation-request.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry