EPA Emissions MCP has well-structured tool definitions with complete input schemas and reasonable descriptions. All 5 tools are explicitly registered with name, description, and inputSchema. However, descriptions lack LLM-optimization guidance (e.g., when to use each tool vs. alternatives), parameter descriptions are minimal, and output schemas are entirely undocumented. No error handling guidance, no examples of what fields are returned, and no documentation of pagination behavior despite limit parameters. Tool naming is clear and verb-based (search_, get_), but composition could be stronger, tri_trends requires either state OR chemical but this mutual-optionality is not clearly documented.

• Document output schemas for all tools. Add to each description: 'Returns: [facility_id, name, city, state, zip, latitude, longitude, industry_type, total_ghg_emissions_mt_co2e, year]' (or equivalent for TRI tools). This enables LLMs to plan downstream calls and extract the right fields.
• Clarify tri_trends required parameters. Either make state and chemical both required, or document the behavior when both are omitted. Add: 'At least one of state or chemical must be provided. If both are omitted, returns trends for all facilities and chemicals.'
• Add pagination guidance to all tools with limit parameter. Document: 'Results are returned in batches. Use limit (1-100, default 20) to control batch size. If more results exist, the response includes a next_cursor field for fetching the next batch.'
• Add error handling examples to tool descriptions. E.g., 'If state is invalid, returns error: "State not found. Valid states: [list]. Did you mean: [suggestions]?" If no facilities match, returns empty list.'
• Standardize state parameter naming. ghg_* tools expect full names ('Texas'), tri_* tools expect abbreviations ('TX'). Rename parameters to state_name vs state_abbr, or accept both formats and document the difference clearly.