mcp-everest has 5 tools with basic schemas and descriptions, but significant gaps limit production readiness. All tools have descriptions (10-200 chars) and input schemas with type definitions, meeting baseline requirements. However, descriptions lack actionable context for LLM selection (e.g., 'Get details' is vague; no guidance on when to use vs. list_database_clusters). Parameters lack constraints (enums, ranges, format hints). Error handling returns generic {'error': str(e)} without recovery guidance. No output schema documentation. The create_database_cluster tool has good parameter coverage with defaults, but allow_unsafe=true as default is a dangerous pattern. Tool naming is clear (verb_noun), but composition could be improved, credentials and components are separate calls when often needed together.

• Add output schema documentation to each tool. Document the structure of returned objects, especially IDs and references needed for chaining (e.g., cluster_id, namespace).
• Enhance error handling: categorize errors as retryable (network timeout, 503) vs. user-fixable (invalid namespace, cluster not found) vs. fatal. Return actionable guidance: 'Cluster not found. Call list_database_clusters(namespace="everest") to see available clusters.'
• Expand tool descriptions to 100-150 chars with context: 'Get full details of a database cluster including status, replicas, and resource usage. Use this after list_database_clusters() to inspect a specific cluster before modifying it.'
• Change allow_unsafe default to false. Require explicit allow_unsafe=true for unsafe configurations. Document what 'unsafe' means in the parameter description.
• Add enums for engine_type (e.g., ['pxc', 'mysql', 'postgresql']) and format constraints for storage_size and memory (e.g., pattern: '^\d+(Mi|Gi)$'). Document in parameter descriptions.

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

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