TinaMCP provides 4 read-only tools with basic descriptions and input schemas. All tools have descriptions (10-100 chars) and parameter types defined. However, descriptions are minimal (avg 80 chars vs baseline 194), lack WHEN/WHY guidance, and omit output schema documentation. No error recovery hints, no pagination support despite list tools, and no tool annotations (readOnlyHint). Path traversal validation is present in code but not documented in parameter descriptions. Tools follow verb_noun naming but descriptions do not explain distinctions between ListCollectionDocuments vs ListDocumentsRecursive clearly enough for LLM selection.

• Add output schema documentation to each tool. E.g., GetTinaDocument returns string (file content); ListCollectionDocuments returns List<string> (filenames). Document expected format (markdown, YAML, JSON).
• Expand descriptions to 100-200 chars with WHEN/WHY. E.g., 'Lists all documents in a collection (non-recursive). Use this for flat collections; use ListDocumentsRecursive for nested subdirectories.'
• Add pagination: limit (default 20, max 100) and offset parameters to list tools. Return total_count and has_more in response to enable agent-driven pagination.
• Enhance error messages with recovery hints. E.g., 'Collection "posts" not found. Available collections: [blog, docs]. Try ListTinaCollections() to discover valid names.'
• Add readOnlyHint annotation to all tools in schema to signal safety and idempotency to agents.

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

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