ClawHub
Back to skill

Security audit

OpenTalk2HTML-NotMD MCP Server

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent HTML-generation MCP server, with expected local file-writing and email-template risks that users should handle deliberately.

Install this if you want an MCP server that can create and modify HTML reports, dashboards, invoices, and similar pages. Use a dedicated workspace or temporary output folder, review file paths before allowing edits, avoid sensitive paths, and confirm what the email tools do before using them with confidential content.

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 (3)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill prominently documents tools that write, patch, and modify local HTML files, including explicit paths like /tmp/report.html, but it does not warn users that these operations change local filesystem contents. In an agent setting, that omission can lead to silent file creation or modification when a user may believe the skill is only generating content conceptually, increasing the risk of unintended local changes or overwrite of sensitive files if unsafe paths are used.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The skill advertises email-related functionality such as render_email and html_to_email without any warning that email content may be transmitted externally or may contain sensitive information. In an AI-agent workflow, this can normalize sending generated or transformed content through email channels without clear consent, creating privacy and data-leakage risk.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The README advertises tools that can write HTML to disk and convert/render HTML for email, but it does not warn users that these actions may expose sensitive data, create persistent artifacts, or send content beyond the local environment. In an AI-agent context, omission of privacy and consent guidance is risky because users may enable the server assuming it is presentation-only, while an agent could persist or distribute sensitive content through files or email workflows.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.