Landing Page Angle Tester

• Outcome: Pass
• Latest audit: May 28, 2026, 1:34 PM
• Version:
  1.0.0

Overview

This is mainly a local landing-page copywriting helper; it contains dormant generic audit code, but the shipped configuration and instructions keep normal use focused on user-provided marketing input.

Reasonable to install for drafting landing-page messaging, with caution about provenance and extra dormant audit code. Use it with explicit marketing input files, choose output paths deliberately, and review generated claims before publishing. Do not point it at sensitive repositories or private documents unless you are comfortable with local report output containing derived content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Description-Behavior Mismatch

High

• Confidence: 98% confidence
• Finding: The script’s behavior substantially exceeds the declared purpose of generating landing-page messaging angles. It can inspect arbitrary files/directories, scan for secrets and risky shell patterns, and audit skill package structure, which creates a clear capability mismatch that can mislead users into granting access to unrelated local data. In the context of a marketing skill, this mismatch increases risk because operators may not expect broad filesystem analysis.

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: The built-in pattern scanning logic searches arbitrary content for secrets, internal URLs, and shell-execution patterns, which is unrelated to landing-page angle testing. Even without exfiltration code, this enables unexpected inspection of sensitive local material and may surface confidential snippets in generated reports or stdout. The declared skill context makes this more suspicious because users would not reasonably expect security scanning behavior from a messaging tool.

Context-Inappropriate Capability

Medium

• Confidence: 94% confidence
• Finding: The directory and package audit functions enumerate arbitrary files, inspect markdown headings, and summarize repository structure outside the skill’s marketing use case. This broad local introspection can expose sensitive project metadata and content to reports, especially when pointed at large or unrelated directories. Because the skill is advertised for landing-page messaging, the expanded filesystem access is less expected and therefore more dangerous.

VirusTotal

VirusTotal findings are pending for this skill version.