Cloud Phone Agent

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed cloud-phone automation helper, but it depends on a powerful external CLI and API key that users should treat carefully.

Install only if you trust the mcporter package and your configured cpc-mcp-server. Use a least-privilege CLOUDPHONE_API_KEY, prefer test accounts and non-production cloud phones, review the exact app action before dispatching tasks, and avoid mcporter full-URL or --stdio modes unless you intentionally want those broader capabilities.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Context-Inappropriate Capability

High
Confidence
93% confidence
Finding
The README explicitly documents arbitrary remote URL invocation and arbitrary local stdio command execution, which are much broader than the skill's stated purpose of calling a specific cloud Android agent. In the context of an agent skill, these capabilities can enable SSRF-like outbound access, data exfiltration, or execution of unintended local programs if the wrapper does not strictly constrain inputs.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The examples normalize direct HTTP access and local stdio execution without any warning about network egress or command execution. In an agent-skill context, that omission increases the chance that operators expose dangerous capabilities without realizing they permit outbound requests and arbitrary local process launch.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal