ClawHub

Miniature Build Video

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed WeryAI miniature-video generation skill that uses an API key and may send approved prompts or images to WeryAI, with no evidence of hidden persistence or destructive behavior.

Install only if you trust WeryAI and the publisher with your WERYAI_API_KEY, prompts, image URLs, and any explicitly uploaded local image files. Use dry-run first, avoid sensitive content, prefer public HTTPS image URLs, and confirm the full expanded prompt and exact image path before any paid run.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Natural-Language Policy Violations

Medium
Confidence
78% confidence
Finding
The skill requires rewriting the user's prompt into a full English production prompt before submission, unless the user already supplied a long prompt and asked not to rewrite. That creates an unnecessary transformation step that can alter user intent, introduce unsafe or policy-violating details, and increase the chance that sensitive information is embellished or forwarded to the third-party provider without clear user consent.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The implicit trigger test uses a broad natural-language request that can activate the skill based on loosely related phrasing rather than clear, bounded criteria. This increases the chance of over-triggering, where the agent invokes the video-generation skill in contexts the user did not explicitly intend, potentially causing unnecessary external API usage, cost, and unexpected handling of media inputs.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal