Two tools with clear verb-noun naming (get_transcript, get_video_metadata) and reasonable descriptions (60-80 chars). Both have documented input schemas with types and descriptions. However, output schemas are completely undocumented, tools return plain text or unstructured JSON without field definitions. Parameter descriptions lack format constraints and validation guidance. No error handling documentation or recovery guidance. Missing tool annotations (readOnlyHint present in risk field but not in schema). Descriptions are functional but minimal; they lack WHEN to use guidance and dependency hints.

• Document output schemas explicitly. For get_transcript, specify: returns {video_id: string, language: string, source: string, segments: [{start: number, end: number, text: string}], text: string}. For get_video_metadata, specify: returns {video_id: string, title: string, channel: string, duration: number, caption_languages: [string]}.
• Convert 'format' parameter to enum constraint: format: {type: 'string', enum: ['text', 'json', 'srt'], default: 'text'} and update description to 'Output format (text, json, or srt).'
• Add language code guidance: 'lang: {type: string, description: "ISO 639-1 language code (e.g., en, es, fr). Defaults to en. Availability depends on video captions.", default: "en"}'.
• Implement structured error responses. Instead of plain text, return {ok: false, error_code: 'VIDEO_NOT_FOUND'|'NO_CAPTIONS'|'INVALID_FORMAT'|'RATE_LIMITED', message: '...', retry_after?: number}. Update tool descriptions to document these error codes.
• Add readOnlyHint: true to both tool schemas to signal they are safe to call repeatedly without side effects.