Three tools with minimal parameter validation and weak descriptions. test_connection and system_diagnostics have empty input schemas ({}), which is correct for no-param tools, but descriptions lack actionable context. download_video_to_cloud has only one parameter (url: string) with a generic description that includes marketing language ('85-95% success rates') rather than operational guidance. No output schemas documented. No error handling guidance. No parameter constraints (enums, patterns, ranges). Tool names are reasonable (verb_noun style) but descriptions fall short of the 50-200 char LLM-optimized baseline.

• Document output schemas for all three tools. E.g., test_connection should return {success: boolean, endpoint: string, bucket: string, message: string}. system_diagnostics should return {platform: string, ffmpeg_available: boolean, yt_dlp_version: string, success_rate: number}. download_video_to_cloud should return {video_id: string, s3_url: string, file_size_bytes: number, platform: string, duration_seconds: number}.
• Rewrite download_video_to_cloud description to: 'Download videos from YouTube, Instagram, TikTok, Facebook, or LinkedIn and store in S3-compatible cloud storage. Returns S3 URL and metadata. Fails gracefully with platform-specific recovery hints.' Remove marketing claims.
• Add parameter constraints to download_video_to_cloud: url should have a description like 'Video URL from YouTube, Instagram, TikTok, Facebook, or LinkedIn. Must be a valid HTTP(S) URL.' Consider adding an enum or pattern validation.
• Enhance error responses with recovery guidance. E.g., if S3 connection fails, return: 'S3 connection failed: [error]. Check: (1) endpoint URL is correct, (2) credentials are valid, (3) bucket exists. Call test_connection to diagnose.'
• Expand test_connection and system_diagnostics descriptions to 80-150 chars each. E.g., test_connection: 'Verify S3 connectivity and display current configuration. Call this first to ensure cloud storage is accessible before downloading videos.' system_diagnostics: 'Check system dependencies (ffmpeg, yt-dlp), platform support, and recent success rates. Use to diagnose download failures.'

Inferred effective spec: <=2025-11-25.

• Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained