ClawHub
Back to skill

Security audit

UCP Gateway - Agent Commerce

Security checks across malware telemetry and agentic risk

Overview

This shopping gateway skill is purpose-aligned and clearly discloses its remote MCP use, local identity files, and checkout confirmation limits.

Install this only if you want agent-assisted shopping through the UCP Gateway. Expect it to contact ucpg.ai and create ./.ucpgateway identity files. Review product, cart, and checkout details before confirming actions, keep the private key local, and enter payment details only on the merchant-hosted checkout page.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill instructs the agent to use network access, local file storage, and shell-based Python helpers, but it does not declare permissions or narrowly scope those capabilities. This creates a trust and transparency gap: an agent/runtime may grant broader access than users expect, and adversarial modifications to helper scripts or local-state handling could lead to unintended file access, command execution, or data exposure.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.