Premium Solidity smart contract vulnerability scanner REST API & MCP server
Four tools with complete schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names follow verb_noun convention (scan_contract, quick_scan_contract, analyze_gas, list_vulnerabilities). Descriptions are present (avg 156 chars) but lack actionable context about when to use each tool vs. alternatives. All tools accept only 'code' parameter with minimal constraint documentation. No output schemas documented, LLMs cannot predict response structure. Error handling returns generic messages without recovery guidance. No tool annotations (readOnlyHint, idempotentHint). Composition is reasonable: four distinct concerns, but quick_scan_contract and scan_contract overlap significantly without clear differentiation guidance.
Analyze Solidity code for gas optimization opportunities. Returns specific suggestions with estimated gas savings.
List all vulnerability types that the scanner can detect, with descriptions and severity levels.
Quick security scan checking only the top 5 critical vulnerability patterns: reentrancy, access control, tx.origin, delegatecall, and unchecked calls.
Perform a full security audit on Solidity smart contract source code. Detects 13 vulnerability classes including reentrancy, access control, integer overflow, and more. Returns a detailed report with risk score, findings by severity, and recommendations.
No output schemas documented. LLMs cannot predict response structure (fields, types, nesting). Responses are JSON-stringified text blobs without type hints.
Parameter 'code' lacks format/length constraints. No guidance on max code size, Solidity version, or expected structure. LLMs may pass invalid or oversized inputs.
scan_contract and quick_scan_contract have overlapping purpose. Descriptions do not explain when to choose one over the other. LLMs may pick wrong tool or call both unnecessarily.
Error messages are generic ('Tool execution failed: ...') without recovery guidance. LLMs cannot self-correct or know whether to retry, ask user, or abandon.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
No tool annotations (readOnlyHint, idempotentHint). All tools are read-only but this is not declared. Agents cannot optimize caching or retry logic.