Scopus MCP has well-structured tool definitions with clear naming (verb_noun pattern) and consistent parameter schemas using Zod. All 5 tools have descriptions and input schemas. However, descriptions are brief (avg ~80 chars, below the 194-char baseline), lack dependency hints and recovery guidance, and output schemas are not documented. Error handling wraps calls but returns generic JSON errors without actionable recovery steps. Parameters lack enum constraints where applicable (e.g., 'sort' accepts free-form strings). No tool annotations (readOnlyHint, etc.) despite all tools being read-only.

• Document output schemas for each tool. E.g., search_scopus returns {results: [{scopus_id, title, authors, year, citationCount}], total: number}. This enables LLMs to extract fields without guessing.
• Enhance error messages with recovery guidance. Instead of 'Unknown error', return 'API rate limit exceeded. Wait 60 seconds before retrying.' or 'Invalid Scopus ID format. Expected format: SCOPUS_ID:12345678.'
• Convert 'sort' parameter to an enum: z.enum(['coverDate', 'relevancy', 'citationCount']).default('coverDate'). This prevents invalid sort orders and self-documents valid options.
• Expand tool descriptions to 150-250 chars. Include: what the tool does, when to use it (vs similar tools), what it returns, and any prerequisites. E.g., 'Search Scopus for documents matching a query. Use this to discover papers before calling get_abstract_details for full metadata. Returns up to 25 results with title, authors, year, and citation count.'
• Add readOnlyHint: true to all tool definitions to signal to agents that these calls are safe to retry and have no side effects.