ClawHub
Back to skill

Security audit

Embodied Ai Weekly

Security checks across malware telemetry and agentic risk

Overview

The skill is coherent for generating and publishing an embodied-AI weekly report, but its publish workflow can directly change a live GitHub Pages repository without a required final review step.

Install only if you are comfortable supervising the publishing step. Use a dedicated GitHub Pages repository or limited deploy key, require the agent to show git status and git diff before any commit or push, and prefer a branch or pull request instead of direct pushes to main.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly instructs cloning a user-specified repository, creating or modifying files, committing, and pushing to the main branch, which can directly alter published GitHub Pages content. Because these are state-changing external actions and the skill does not prominently require explicit user confirmation immediately before publish, an agent could perform unintended publication or overwrite existing site content.

Missing User Warnings

Low
Confidence
82% confidence
Finding
The skill directs bulk downloading of images from arXiv pages and screenshots of GitHub/project pages into the workspace, creating many files and potentially republishing third-party visual content. Without warnings about storage impact, attribution, licensing, and source-verification responsibilities, this can lead to copyright/compliance issues and unexpected filesystem changes.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.