MCP server providing regulatory compliance tools for DeFi agents, including MiCA authorization checks, risk assessment, macro conditions evaluation, and on-chain evidence anchoring to Arbitrum
Server has 6 tools with schemas and descriptions present, but quality is inconsistent. Tool names lack action verbs (compliance_preflight, risk_assessment, recession_risk, audit_log, anchorEvidence, verifyEvidence). Most descriptions are adequate (100-200 chars) but lack LLM-optimized guidance on WHEN to use each tool vs alternatives. Parameter descriptions exist but are minimal (10-30 chars). Critical gap: no output schemas documented anywhere, LLMs cannot plan downstream calls or extract structured data. audit_log and anchorEvidence are write operations but lack confirmation/dry-run patterns. No error handling guidance visible in tool definitions.
Write compliance decision hashes to Arbitrum for permanent verification. Every PASS/WARN/BLOCK decision gets an on-chain proof.
Create tamper-proof audit entry with evidence chain for trade decisions
Check MiCA authorization and evidence quality for a token before executing trades or swaps
Check macro environment and recession risk conditions
Get 7-signal risk score and verdict for a token
Verify compliance evidence on-chain by retrieving and validating transaction data from Arbitrum
No output schemas documented. LLMs cannot infer what fields to expect from tool responses, blocking downstream tool chaining and data extraction.
Tool names lack action verbs. 'compliance_preflight' should be 'check_compliance' or 'verify_compliance'; 'risk_assessment' should be 'get_risk_score' or 'assess_risk'. Verb-first naming helps LLMs infer intent from the name alone.
Write operations (audit_log, anchorEvidence) lack confirmation or dry-run patterns. Agents may accidentally log false decisions or anchor incorrect evidence on-chain without a safety gate.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
Parameter descriptions are minimal (10-30 chars). E.g., 'Token symbol' lacks context on format (uppercase? with chain?), valid range, or when to use. Descriptions should be 50-150 chars with format/constraint hints.
No error handling guidance in tool definitions. If compliance_preflight returns BLOCK, what should the agent do? Retry? Ask the user? Abort? Missing recovery guidance forces LLMs to guess.