Two tools with basic verb-noun naming (search_law, list_laws) and Thai-language descriptions. Both have input schemas with types and defaults. However, descriptions are minimal (under 100 chars), lack WHEN/WHY context, and parameter descriptions are sparse. No output schema documentation. No error handling guidance. Schemas present but incomplete, missing validation constraints, ranges, and recovery hints. Follows ~20% of the 54 patterns.

• Add English descriptions (50-150 chars) explaining WHAT, WHEN, and WHY. Example: 'Search Thai law sections by semantic similarity. Use when user asks about legal rights, obligations, or penalties. Returns ranked sections with relevance scores (0-1).'
• Document output schema in tool docstring. Example: 'Returns: list[{section: str, law_name: str, text: str, score: float}]'
• Add constraints to parameters: query (1-500 chars, Thai or English), n_results (1-50, default 5). Include in description: 'n_results: Number of results to return (1-50, default 5). Higher values increase latency.'
• Implement error handling with recovery hints. Catch ChromaDB connection errors and return: 'ChromaDB unavailable. Retry in 5 seconds or check CHROMA_HOST/CHROMA_PORT configuration.'
• Add tool annotations (readOnlyHint: true) to both tools since they only read data and have no side effects.