ISE-MCP has 10 tools with basic descriptions and input schemas, but significant quality gaps limit production readiness. All tools have descriptions (10-100+ chars), and most have input schemas with type definitions. However, descriptions are generic and lack LLM-optimization guidance (e.g., 'List all files in the root folder or specified folder' doesn't explain WHEN to use it vs similar tools). Parameter descriptions are minimal or absent. Output schemas are documented only in JSDoc comments, not in the tool registration itself. No error handling guidance, no pagination support despite list tools, and no tool annotations (readOnlyHint, etc.). The server uses workers-mcp HTTP transport (good), but lacks modern MCP patterns like structured output documentation, per-request _meta logLevel, and MRTR support.

• Add pagination parameters (limit, offset) to all list tools. Document max result limits (e.g., 'Returns up to 50 items; use offset for pagination'). Include total_count in responses.
• Move output schema definitions from JSDoc into explicit JSON Schema objects in tool registration. Document field types, nesting, and required fields so LLMs can plan downstream calls.
• Rewrite descriptions to be LLM-optimized (50-200 chars). Include WHEN to use each tool: 'List all files in the current folder. Use this to discover file IDs before calling parseFile(). Returns {name, id, mimeType}.'
• Add tool annotations to the schema: set readOnlyHint=true for all tools. This signals to MCP clients that these tools are safe to call without confirmation.
• Add error handling guidance to parseFile and parseAndAnalyzeFile: 'If parsing fails, check isFileTypeSupported(mimeType) first. Supported types: PDF, TXT, CSV, JSON, Markdown, Google Docs/Sheets/Slides.'