ClawHub
Back to skill

Security audit

Agent Bounty Scanner

Security checks across malware telemetry and agentic risk

Overview

This skill transparently runs the ACP marketplace browser command to list and rank bounty opportunities, with no evidence of hidden data theft, persistence, or destructive behavior.

Install this only if you also trust the separate `virtuals-protocol-acp` skill or CLI, because that component performs the actual marketplace interaction. Prefer passing a known ACP executable path instead of relying blindly on PATH, and review ACP permissions before using this in autonomous workflows.

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
91% confidence
Finding
The skill documentation indicates it invokes the external `acp` command via subprocess, which is a shell-capable operation, but the skill declares no permissions. Even if argument lists are used safely, undeclared command execution expands the skill's effective privilege surface and can mislead users or platforms about what the skill is allowed to do, especially since it relies on PATH resolution (`acp_command="acp"`) which may execute an unexpected binary in a compromised environment.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.