Server has 4 well-named search/retrieval tools with complete input schemas and reasonable descriptions. All tools follow verb_noun naming (search_*, get_*). Schemas are properly typed with descriptions for all parameters. However, output schemas are not documented, LLMs cannot predict response structure for downstream chaining. Error handling is minimal (generic RuntimeError). Parameter descriptions are functional but lack constraint details (e.g., no min/max for page/limit, no format specs for dates). Tool descriptions average ~150 chars, within baseline range, but lack WHEN-to-use guidance and dependency hints between tools.

• Document output schemas for all tools. Example: 'Returns {hits: [{id, title, description, modified, ...}], total: int, facets: {...}, scroll_id?: string}'. Include field types and nesting.
• Add constraint descriptions to numeric/date parameters. E.g., 'page: Zero-based page number (0 - 1000)' and 'min_date: ISO 8601 date string (YYYY-MM-DD), e.g., 2020-01-01'.
• Enhance tool descriptions with WHEN-to-use guidance. E.g., 'Use piveau_search_datasets for keyword search with basic filters. Use piveau_search_datasets_advanced for date ranges, geographic bounding boxes, or field-level scoring.'
• Add error recovery hints to tool descriptions. E.g., 'If no results found, try broadening filters or using piveau_search_datasets_advanced with different facets.'
• Document pagination behavior: 'Returns up to `limit` results (default 10, max 100). Check `total` field to determine if more pages exist. Use `page` parameter to fetch next page.'