14 tools with consistent naming (verb_noun pattern: get_, search_, list_, submit_, cancel_). All tools have descriptions (10-60 chars, below ideal 50-200 baseline). Input schemas present for all tools with parameter types and descriptions. However, descriptions are terse and lack context on WHEN to use each tool, dependencies, or error recovery guidance. Output schemas not documented. No tool annotations (readOnlyHint/destructiveHint). Error handling returns structured JSON but lacks actionable recovery guidance for LLMs.

• Expand tool descriptions to 50-200 characters. Include WHEN to use, prerequisites, and how it differs from similar tools. E.g., 'Get the full transcript for a YouTube video. Use this after search_youtube to retrieve the text. Returns transcript text, language, and metadata. Requires video_id.'
• Document output schemas for all tools. Add a 'returns' section to each tool definition listing field names, types, and descriptions. E.g., get_transcript returns {transcript: string, language: string, kind: string, metadata: object}.
• Add tool annotations: mark all read-only tools with readOnlyHint=true; mark submit_transcripts_job and cancel_transcripts_job with destructiveHint=true; mark idempotent tools with idempotentHint=true.
• Enhance error messages with recovery guidance. Instead of 'HTTP 503', return 'Service temporarily unavailable. Retry in 30 seconds or check https://status.transcriptout.com.' For 'video not found', suggest 'Try search_youtube to find the correct video ID.'
• Document pagination for list/search tools. Add limit and offset parameters to list_channel_videos, list_playlist_videos, search_youtube, search_channel_videos, search_playlist_videos. Return total_count and next_offset in responses.