Server has well-structured tool definitions with complete schemas and reasonable descriptions. All 5 tools have input schemas with type definitions and parameter descriptions. However, descriptions are generic and lack LLM-optimized guidance on WHEN to use each tool vs alternatives. Output schemas are not documented. Error handling is minimal. Parameter descriptions lack format constraints and validation rules. No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications.

• Add output schema documentation to each tool. Example for summarize_video: 'Returns {summary: string, duration_seconds: number, key_timestamps: [{time: number, label: string}]}'. This enables LLMs to chain tools and extract structured data.
• Add tool annotations to all tools. Use readOnlyHint: true for summarize_video, ask_about_video, get_video_timestamps. Use destructiveHint: true for extract_screenshots and extract_frames (they write to disk).
• Expand descriptions to 100-150 chars with WHEN guidance. Example: 'Summarize a YouTube video. Use this for quick overviews; use ask_about_video for specific questions. Returns text summary with optional timestamps.'
• Add enum constraint for 'focus' parameter in extract_screenshots and get_video_timestamps. Example: enum: ['product_demos', 'code_examples', 'diagrams', 'text_overlays', 'general_key_moments'].
• Add error recovery hints to extract_screenshots and extract_frames descriptions. Example: 'If extraction fails, the video may be private, age-restricted, or too long. Ask the user to verify the URL or try a lower resolution.'