Unitree Hermes Colab

• Outcome: Pass
• Latest audit: Jun 6, 2026 at 7:28 AM UTC
• Version:
  0.1.0

Overview

This skill is a clearly scoped Unitree/Hermes Colab helper with explicit safety boundaries and no evidence of hidden, destructive, or deceptive behavior.

Install this only if you want Codex help building or reviewing the Unitree Hermes Colab workflow. Review generated notebooks before running them, keep Hermes execution opt-in, and provide model API keys only when you intend to run provider-backed analysis.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The skill enables implicit invocation without any visible trigger constraints, so the agent may activate this workflow in broader contexts than intended. Because this skill can drive package installation and Colab-based analysis workflows, ambiguous activation increases the chance of unintended execution, prompt-scope confusion, or use outside its stated safety-gated purpose.

VirusTotal

65/65 vendors flagged this skill as clean.