MCP server that calculates sales tax rates and nexus determination for different US jurisdictions based on cart subtotal and shipping address
Single tool 'execute_skill_action' has a reasonable description (89 chars) and input schema with typed parameters, but critical gaps prevent higher scoring. The schema lacks output documentation entirely, LLMs cannot plan downstream calls without knowing what fields are returned. Parameter descriptions are present but minimal (e.g., 'Shipping address with city, state, and zip' lacks format/validation guidance). No error handling guidance, no recovery paths, and no indication of what the tool actually returns beyond the description. The tool name 'execute_skill_action' is generic and violates verb_noun clarity, 'calculate_sales_tax' would be more specific. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite being READ_ONLY. The implementation in client.py shows hardcoded tax rates for only 3 states with no extensibility or error handling for unmapped jurisdictions.
Execute sales tax calculation and nexus determination for a given cart subtotal and shipping address
No output schema documented. LLMs cannot infer what fields the tool returns (tax_due, nexus_status, etc.), forcing them to guess or make wasteful follow-up calls.
Tool name 'execute_skill_action' is generic and violates verb_noun convention. Should be 'calculate_sales_tax' or 'calculate_sales_tax_and_nexus' to clearly signal intent.
Parameter descriptions lack format/validation guidance. 'State abbreviation (e.g., TX, CA, NY)' should specify enum constraint or pattern. 'ZIP code' lacks format (5 digits, optional +4, etc.).
No error handling guidance. Implementation silently defaults to TX for unmapped states, LLM receives no indication of fallback behavior or how to handle unsupported jurisdictions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 38 | 2026-07-28+ | v2 |
Tool marked READ_ONLY but lacks tool annotations (readOnlyHint). MCP spec allows explicit annotation to signal safety to clients.