ClawHub
Back to skill

Security audit

Train Robotic AI Models using Qualia

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Qualia API wrapper for launching and managing robot-model training jobs, with real account and credit impact that users should control carefully.

Install only if you trust Qualia and are comfortable giving the agent a Qualia API key that can view credits, create projects, launch paid training jobs, cancel jobs, and delete projects. Require explicit confirmation before any paid training launch or project deletion, and prefer a limited or revocable API key if Qualia supports it.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill explicitly requires an API key and instructs users to run a Python CLI that makes authenticated network requests, but it does not declare permissions despite clearly using env, network, and shell capabilities. This creates a transparency and trust gap: an agent or user may invoke a capability-bearing skill without an explicit permission model, increasing the risk of unintended secret use, external data transmission, or command execution.

Vague Triggers

Medium
Confidence
75% confidence
Finding
The guidance says that when someone asks to train a model, the agent should gather missing details, choose defaults, create or reuse projects, and prepare to launch training jobs. In an agentic environment, this broad invocation guidance can cause over-eager use of a networked, cost-incurring skill with access to an API key, potentially triggering external actions or spending credits without sufficiently explicit user confirmation.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The CLI exposes a destructive project deletion operation without any interactive confirmation, dry-run, or force/confirm safeguard. In an agent or automation context, a mistaken argument, copied ID, or prompt-induced command could irreversibly delete a project, making accidental destructive actions materially more likely.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.