Three tools with complete schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names lack action verbs (datasets, query, metadata are nouns). Descriptions are adequate (100-200 chars) but lack WHEN/WHY guidance. Parameter descriptions exist but omit constraints, ranges, and format details. No output schema documented. Error handling is minimal. Composition is sound, tools chain logically (datasets → metadata/query). Security is acceptable (read-only, no secrets exposed). Naming could be clearer: 'search_datasets', 'run_query', 'get_metadata' would better signal intent to LLMs.

• Rename tools to action verbs: 'search_datasets', 'run_query', 'get_metadata'. This signals intent to LLMs and aligns with production baselines (90% of A+ tools start with verbs).
• Document output schemas in tool descriptions. E.g., 'Returns: {results: [{resource_id, name, description, category, updated}], total_count}' for datasets.
• Add format/constraint details to parameter descriptions. E.g., 'where: SoQL WHERE clause (e.g., "year >= 2020 AND status = \'Active\'"). Do not include leading $where.'
• Add error recovery guidance. E.g., 'If resource_id is invalid, call search_datasets() first to find valid IDs.' This guides LLM retry logic.
• Document pagination behavior in 'datasets' description: 'Returns up to `limit` results (default 20, max 100). Use `offset` to paginate. Response includes total_count for planning subsequent calls.'