ClawHub
Back to skill

Security audit

Luckin Coffee

Security checks across malware telemetry and agentic risk

Overview

The skill is for Luckin Coffee account actions, but its read-only description conflicts with real order creation and cancellation capabilities.

Review this skill carefully before installing. It can interact with a connected Luckin Coffee account, including placing orders that may require payment and canceling orders. Only use it if you are comfortable with the oo/OOMOL connector path and make sure the agent asks for explicit confirmation before any createOrder or cancelOrder action.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

High
Confidence
98% confidence
Finding
The manifest and description claim the skill is for 'searching and reading data', but the documented actions include real order creation and cancellation. This mismatch can cause an orchestrator or user to treat the skill as read-only and invoke it in situations where writes should be blocked or require stronger confirmation, creating a real risk of unauthorized purchases or cancellations.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The document says state-changing actions will be marked with '[write]' or '[destructive]', but the action list does not actually apply those tags to createOrder and cancelOrder. This inconsistency can mislead downstream agents or reviewers into classifying those actions as safe reads, increasing the chance of accidental order placement or cancellation.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger language says to use this skill for 'ANY Luckin Coffee request', which is overly broad for a skill that includes write-capable operations. Broad routing increases the chance the skill is invoked automatically for casual informational requests, potentially exposing users to unintended state-changing pathways if the agent mishandles intent or confirmation.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.