Python client for the MVR API relational-readiness layer
This MCP server wraps a Python API client for the MVR (Minimum Viable Relationships) service. Tools are present with descriptions and some parameter schemas, but critical gaps undermine quality. Most tool descriptions are under 100 chars and lack actionable context (e.g., 'Verify API authentication status' does not explain when to call it or what to do with the result). Parameter schemas are incomplete: many complex objects like 'subject', 'market_scope', and 'evidence_pack' are typed as generic 'object' with minimal field documentation. No output schemas are documented. Error handling is basic (MVRApiError exists but lacks recovery guidance). Tool naming follows verb_noun convention (auth_check, entity_resolve) but some names are vague (context_compile, decision_check). No tool annotations (readOnlyHint, destructiveHint, idempotentHint) are present. The server appears to be a thin wrapper around an external API without agent-specific optimization.
Verify API authentication status
Retrieve API capabilities and supported features
Compile context from evidence and subject information
Check decision readiness based on compiled context
Resolve an entity name to canonical form with optional country filtering
Evaluate evidence completeness for a subject in a market scope
Complex object parameters (subject, market_scope, evidence_pack) lack field-level documentation. Typed as generic 'object' with only high-level descriptions. LLMs cannot infer required fields, types, or constraints.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields. Responses are opaque.
Tool descriptions are too brief (35 - 50 chars) and lack context. 'Verify API authentication status' does not explain when to call it, what success/failure means, or what to do next. Descriptions should be 50 - 200 chars with actionable guidance.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
Start the bounded MVR workflow; this response is never a verdict.
Check API health status
Retrieve the model card documentation
Get recommended inputs for a given endpoint and entity archetype
Get remediation path recommendations
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) present. LLMs cannot determine which tools are safe to retry or have side effects.
Error handling is minimal. MVRApiError exists but provides no recovery guidance. Errors should categorize as retryable/user-fixable/fatal and suggest next steps.