Single tool with minimal schema and description. Tool name 'chunk_document' lacks action verb prefix (should be 'partition_document' or 'split_document'). Description is 67 chars, acceptable but generic. Input schema present but lacks parameter descriptions entirely. No output schema documented. No error handling guidance. Missing 90% of production-grade patterns.

• Rename tool to 'partition_document' or 'split_document' to start with a clear action verb. This helps LLMs infer intent from the name alone.
• Add descriptions to all input parameters: 'text' (required, the document to partition), 'max_words' (optional, default 40, range 10-500, max words per chunk).
• Document the output schema in the tool description: 'Returns array of objects with chunk_index (int), word_count (int), content (string), chunk_hash (string).'
• Expand tool description to 100-150 chars: 'Partitions long documents into overlapping sentence-aware chunks without breaking grammatical boundaries. Use this to prepare text for embedding or summarization. Returns chunks with word counts and content hashes for deduplication.'
• Add error handling: validate text is non-empty, max_words is 10-500, and return clear errors like 'Text cannot be empty' or 'max_words must be 10-500, got {value}'.