PyMemSim-MCP has 3 tools with READ_ONLY risk classification. Tool names follow verb_noun pattern (simulate_*, check_*, analyze_*), which is good. However, descriptions are present but lack depth, they state WHAT but not WHEN or WHY to use each tool, and lack actionable guidance for LLMs. No input parameter descriptions are visible in the source code; tool signatures show only generic 'request' objects without field-level documentation. Output schemas are not documented. Error handling and recovery guidance are absent. The server exposes domain-specific simulation logic but fails to meet production-grade tool definition standards.

• Add explicit JSON Schema definitions for all tool input parameters. For simulate_gas_hfm, document GasHFMSimulationRequest fields (types, required, constraints, descriptions). Use enums for categorical inputs (e.g., phase type, unit system).
• Expand tool descriptions to 100 - 200 characters. Include: (1) what the tool does, (2) when to use it vs. alternatives, (3) expected input format, (4) what the output contains. Example: 'Simulate gas-phase hollow-fiber membrane performance. Use this after validating thermodynamic data with check_yaml_reference. Returns flow rates, permeances, and pressure drops as JSON.'
• Document output schemas for each tool. Specify return type (object/array), field names, types, and units. For simulate_gas_hfm, describe the structure of simulation results (e.g., {results: {flow_rate: number, unit: string}, metadata: {solver_time_ms: number}}).
• Add error handling with actionable recovery guidance. Example: 'YAML validation failed: missing CO2-g permeance data. Call check_yaml_reference() to identify missing fields, then update the reference file.'
• Add parameter descriptions to all tool inputs. For each field in GasHFMSimulationRequest, specify: type, valid range/enum, units, and whether it's required. Example: 'feed_temperature (number, required): inlet gas temperature in Kelvin (273.15 - 500).'.