Three tools with moderate naming and descriptions, but critical schema gaps. Tool names lack action verbs (convert, ingest, doctor are weak). Descriptions exist but are generic (194-char baseline; these average ~120 chars). Input schemas are partially visible but lack comprehensive parameter descriptions and type constraints. The 'convert' tool has a complex options object with enum constraints (mode: fast|balanced|thorough), but most parameter descriptions are minimal (10-20 chars). 'ingest' and 'doctor' schemas are sparse. No output schemas documented. Error handling and recovery guidance absent. Composition is reasonable (three distinct concerns), but parameter naming could be more explicit (e.g., 'input' is vague; 'source_path' or 'file_path' would be clearer).

• Rename tools to action verbs: 'convert' → 'convert_to_markdown', 'ingest' → 'ingest_folder', 'doctor' → 'check_system_dependencies'. This signals intent and disambiguates from similar tools.
• Document output schemas for all three tools. E.g., convert returns {title, source, sourceType, sections[], metadata{extraction, extraction_status, kind}}. This enables LLMs to chain calls and extract fields.
• Expand parameter descriptions to 50-100 chars. E.g., 'summary: Include a brief summary of the content at the top of the markdown output (useful for long documents)' instead of 'Include summary in output'.
• Add error handling guidance. E.g., convert description: 'If conversion fails, call check_system_dependencies to verify required tools (tesseract, ffmpeg, etc.) are installed.' This guides recovery.
• For 'doctor' tool, document what it returns: a JSON object with {tesseract: bool, pdftotext: bool, ffmpeg: bool, ...} and explain how to interpret each field.