ClawHub

claw-markdown-gen

Security checks across malware telemetry and agentic risk

Overview

This skill coherently generates rewritten webpage articles and optional AI image links, with no evidence of hidden exfiltration or destructive behavior.

Install only if you are comfortable sending the selected webpage article text and image context to the agent and, in heavy mode, sending generated image prompts to your configured image API. Avoid using it on private dashboards, account pages, or confidential documents unless that data may leave the browser workflow.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The security note claims the generated output contains no executable code or scripts, but the same document explicitly permits raw SVG and Mermaid blocks. In many renderers, SVG can carry active content and Mermaid may introduce rendering-time risks, so this statement is misleading and may cause unsafe downstream rendering assumptions.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The manifest frames the skill as content generation from webpage text, but the workflow additionally supports remote image generation with user credentials and replacement of placeholders with remote URLs. This under-disclosure increases the chance that sensitive article content or prompts are sent off-platform without informed user approval.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The document explicitly states that `context` is extracted from the page DOM and embedded into API requests as plain text. That can transmit surrounding page content to a remote service without clear user notice, consent, or data-minimization controls, creating a real privacy and potential data-leak risk if the captured text includes sensitive on-page information.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal