ClawHub
Back to skill

Security audit

AI Product Launch Monitor

Security checks across malware telemetry and agentic risk

Overview

This skill does what it says: it fetches public AI launch feeds, enriches results with web search, optionally screenshots pages, and writes local reports.

Install only if you are comfortable with outbound requests to RSS feeds, Brave or DuckDuckGo search, and product pages rendered in headless Chromium. Use a dedicated output directory, use --no-screenshots or --no-search in restricted environments, and avoid untrusted custom feeds because their links may be opened by the browser screenshot step.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill clearly instructs execution of a Python pipeline that uses environment variables, writes files, and performs network access, yet the skill metadata shown does not declare those capabilities or permissions. This is dangerous because an agent or operator may invoke it without understanding that it will reach external sites, scrape content, and create artifacts on disk, reducing informed consent and weakening permission enforcement.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill description presents the pipeline as a simple monitoring command but does not warn that it performs external web requests, HTML scraping, and headless-browser screenshot capture against third-party sites. That omission is risky because these actions can trigger privacy, compliance, or operational issues in restricted environments, and users may not realize they are authorizing outbound traffic and automated browsing.

Missing User Warnings

Low
Confidence
86% confidence
Finding
The skill does not prominently warn that it writes markdown reports, JSON output, and screenshot images to disk. This can create unintended data retention, overwrite workspace contents, or expose scraped information to other tools or users if the output path is not carefully controlled.

Missing User Warnings

Medium
Confidence
82% confidence
Finding
The script automatically visits third-party URLs from RSS feeds and saves screenshots locally without an explicit consent or warning step. In an agent-skill context, this can create privacy and safety issues because untrusted pages are fetched and rendered in a browser, potentially accessing sensitive internal URLs if feeds are attacker-influenced or if the user did not expect active browsing and file creation.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.