ClawHub
Back to skill

Security audit

seedance2.0

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Seedance video-generation helper, with normal cautions around API keys, external media generation, and local video files.

Install only if you trust the publisher and the Volcengine partner/application link. Use a restricted ARK_API_KEY, keep ARK_BASE_URL pointed at a trusted endpoint, avoid sensitive prompts or private media URLs, and expect generated videos to be written to local disk.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
87% confidence
Finding
The skill declares access to environment variables and network-capable behavior but does not declare corresponding permissions. This weakens transparency and informed consent for users, especially because API keys are read from the environment and remote services are contacted, creating avoidable risk if the skill is trusted more broadly than intended.

Tp4

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding
The documented purpose does not fully match the described/implemented behavior: it omits local file download/write behavior and additional media input handling while claiming support for text-to-image that is not actually implemented. Description-behavior mismatches are dangerous because they prevent users from accurately understanding what the skill can do, including writing files locally and handling more sensitive user-supplied media than expected.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill states that generated videos are automatically downloaded to local storage without an explicit warning, opt-in, or indication of destination/path handling. Silent file writes can surprise users, consume disk space, overwrite files if poorly implemented, and create privacy or safety issues if generated content is stored on shared systems.

Unpinned Dependencies

Low
Category
Supply Chain
Content
"author": "Zoe",
  "license": "MIT",
  "dependencies": {
    "dotenv": "^16.4.5"
  },
  "openclaw": {
    "install": [
Confidence
91% confidence
Finding
"dotenv": "^16.4.5"

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.env_credential_access

Environment variable access combined with network send.

Critical
Code
suspicious.env_credential_access
Location
seedance2.0.ts:53