ClawHub
Back to skill

Security audit

xCloud Agent Skills

Security checks across malware telemetry and agentic risk

Overview

The skill is a disclosed xCloud hosting operations integration with expected API access and credential setup guidance, but users should treat it as a powerful account-management tool.

Install only if you intend to let an agent manage your xCloud hosting account. Prefer the OAuth MCP connector, grant read-only access unless writes are needed, use the narrowest token scopes for REST fallback, and confirm destructive actions carefully.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • 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
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill advertises and documents shell-based execution (`bash`, `curl`, `jq`, and a shared shell wrapper) but does not declare corresponding permissions. That mismatch can prevent users or the agent platform from accurately understanding the skill's execution and network capabilities, reducing security transparency and increasing the risk of unintended command or API invocation.

Session Persistence

Medium
Category
Rogue Agent
Content
cat >&2 <<'EOF'
error: XCLOUD_API_TOKEN is not set.

Step 1 — Create an API token in xCloud:
  xCloud dashboard -> Profile -> API Tokens -> Generate New Token
  -> choose the scopes you need (e.g. read:servers) -> copy it (shown only once).
Confidence
71% confidence
Finding
The script explicitly instructs users to store a long-lived API token persistently in ~/.claude/settings.json. Persisting bearer tokens in agent configuration increases the blast radius of local compromise, accidental file disclosure, backups, or overbroad tool access, especially in an agent-skill ecosystem that may expose local files to other components.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.