RunContext MCP provides 8 read-only semantic metadata tools with basic descriptions and minimal input schemas. Tool names follow verb_noun convention (search, explain, validate, tier, golden_query, guardrails, list_products, get_product), which is good. However, descriptions are terse (10-50 chars), parameter descriptions are missing or minimal, and output schemas are not documented. Most tools accept a single string parameter with no validation constraints, enums, or format guidance. No error handling guidance visible. The server is functional but lacks the depth expected of production-grade agent tools.

• Expand tool descriptions to 150-250 characters. Include WHAT the tool does, WHEN to use it (vs similar tools), and what it returns. Example: 'Search semantic metadata across models, columns, and glossary terms. Use this to discover available entities before calling explain(). Returns matching entities with type and description.'
• Add parameter descriptions for all string inputs. For 'entity' parameter: 'Entity identifier: model name (e.g., customers), column path (e.g., customers.email), or metric name (e.g., revenue_total). Use explain() to discover available entities.'
• Document output schemas for each tool. Specify returned fields, types, and whether results are paginated. Example for search: 'Returns array of {type: string, name: string, description: string, score: number}. Limited to 50 results; use offset parameter for pagination.'
• Add input validation constraints. Use enums for known values (e.g., tier returns Bronze|Silver|Gold). Add length limits and format hints for string parameters. Example: 'model: string (1-100 chars, alphanumeric + underscore)'.
• Include error recovery guidance in descriptions. Example: 'If model not found, call list_products() to see available models. If validation fails, check that model exists and has required metadata fields.'