Five tools with clear verb-noun naming and detailed descriptions (avg 180 chars, within baseline 194). All tools have input schemas with Zod validation and type definitions. However, output schemas are undocumented, responses are inferred from code (text + image content arrays) but never formally declared. Parameter descriptions are present but lack constraint details (ranges, enums, format specs). Error handling returns toolError() calls but lacks recovery guidance or error classification. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations.

• Add formal output schema documentation for each tool. E.g., get_video_info returns {title: string, channel: string, duration: number, uploadDate: string, viewCount: number, description: string}. Declare this in a schema field or docstring.
• Add min/max constraints to numeric parameters: width ∈ [320, 1920], max_scenes ∈ [1, 50], columns ∈ [1, 6], segment_duration ∈ [30, 600], max_frames ∈ [1, 100]. Document in parameter descriptions.
• Enhance error messages with recovery guidance. E.g., 'Failed to get transcript: language not available. Try language="en" or call get_video_info first to check available subtitles.'
• Add readOnlyHint: true to all five tools via tool annotations. This signals to agents that these tools are safe to call without side effects.
• Clarify watch_video pagination: document that segment_duration controls how much video to process per call, max_frames controls sampling density within that segment, and the response includes next_start_time for continuation.