Loading ToolBench
Flags checkout coupon code usage velocities when they spike above baseline standard limits.
This server exposes a single tool 'audit_redemption' with a reasonable input schema and parameter descriptions, but lacks critical production-grade qualities. The tool name is unclear (audit_redemption is passive/unclear about what happens), the tool description is moderately helpful but generic, and there is NO documented output schema. The implementation is a pure client library with no MCP server transport layer visible, it appears to be a Python SDK, not an MCP server at all. No evidence of MCP protocol implementation, tool registration, error handling patterns, or security controls. The source shows only a raw Python class, not a working MCP server.
Audits coupon code redemption velocity and detects anomalies when daily redemption counts spike above baseline standard limits.
No MCP Server Implementation Visible, Source shows only a Python SDK client class, not an MCP server. No tool registration, message handling, or protocol endpoints are evident.
No Output Schema Documented, The audit_redemption tool returns a dict with fields (coupon_code, anomaly_detected, risk_level, velocity_ratio), but this structure is not formally declared or documented in the tool definition.
Passive/Ambiguous Tool Name, 'audit_redemption' does not start with a clear action verb. It reads as 'perform an audit of redemption', but the intent is unclear (is it reporting, checking, flagging, or analyzing?). A verb-noun pattern like 'detect_redemption_spike' or 'check_coupon_velocity' would be clearer.
No Error Handling or Recovery Guidance, The implementation has no explicit error handling. What if daily_redeemed_count is negative? What if average_daily_redemptions is zero (there is a max(1.0, ...) guard, but it is not documented or part of the tool contract)?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 34 | 2026-07-28+ | v2 |
Tool Description Lacks Specificity, The description 'Audits coupon code redemption velocity and detects anomalies when daily redemption counts spike above baseline standard limits' is generic and does not state the 3x threshold rule, what 'anomaly_detected' actually means, or how risk_level is computed.
Missing Input Validation Documentation, Parameters lack constraint documentation. The description does not specify: What is the valid range for daily_redeemed_count (must be ≥0)? For average_daily_redemptions (must be >0, or defaults to 1.0)? Are string coupon codes case-sensitive?