ff-kit exposes 5 FFmpeg tools with complete input schemas and reasonable descriptions. All tools follow verb_noun naming (ffkit_clip, ffkit_merge, etc.). Schemas include 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, no recovery guidance or categorization. Parameter descriptions lack format constraints (e.g., time format examples are vague). No tool annotations (destructiveHint, idempotentHint) despite all tools being write operations. Missing documentation of mutually exclusive parameters (e.g., end vs. duration in ffkit_clip).

• Add output schema documentation for each tool. E.g., ffkit_clip should document: {status: 'success'|'error', output_path?: string, duration_seconds?: number, error_message?: string}.
• Expand tool descriptions to 150-250 chars. Include WHEN to use (e.g., 'Use to extract audio for speech-to-text pipelines'), prerequisites ('Requires ffmpeg binary'), and side effects ('Creates new file; does not modify input').
• Add tool annotations in MCP response. Mark all 5 tools with destructiveHint=true. Mark ffkit_extract_audio and ffkit_transcode with idempotentHint=true if re-running with same inputs produces same output.
• Document parameter constraints in descriptions. E.g., 'start' in ffkit_clip: 'Start timestamp in HH:MM:SS.ms format (e.g., 00:01:30) or seconds (e.g., 90). Must be >= 0 and < file duration.'
• Implement error categorization and recovery hints. Return {status: 'error', code: 'file_not_found'|'invalid_codec'|'timeout', message: '...', suggestion: 'Try checking file path' or 'Use codec libx264 instead'}.