Security audit

SP10 SKILL

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed product-video scripting helper with minor prompt-scope issues but no hidden code, credential access, persistence, or destructive behavior.

Safe to install for drafting product-video scripts, but review generated handoff prompts before sending product images to an external video tool, especially if you require strict no on-screen text.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Intent-Code Divergence

Medium

Confidence: 96% confidence
Finding: The file establishes a strict no-text/no-subtitles policy throughout the handoff, but the regeneration guidance later suggests using 'more restrained text overlay', which directly conflicts with that rule. In an agent pipeline, this inconsistency can cause downstream prompt generation to reintroduce on-screen text that the user explicitly wanted avoided, leading to policy drift and incorrect or noncompliant outputs.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The manifest description contains very broad trigger language covering many product-video related intents and platforms, which can cause the orchestration layer to invoke this skill in situations beyond its narrow intended use. Over-broad invocation increases the chance of misrouting user requests, suppressing safer or more appropriate skills, and unexpectedly steering users toward scripted commerce/video workflows.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.