ClawHub

Gpt Analyzer

Security checks across malware telemetry and agentic risk

Overview

This skill appears safe to install, but its GPT/model-detection output should be treated as a rough heuristic rather than proof.

Use this skill only as an advisory signal for GPT-like writing patterns. Do not rely on its confidence score or detected model for moderation, academic, employment, or other consequential decisions without independent evidence. The exposed detectVersion/checkWatermarks options should not be assumed to add real analysis behavior in this version.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The skill advertises `detectVersion` and `checkWatermarks` as configuration controls, but neither option affects execution. This creates a security-relevant integrity problem: callers may believe protections or analysis modes are enabled when they are silently ignored, leading to incorrect trust decisions and misleading downstream automation.

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The manifest claims model fingerprinting and version identification, but the implementation only performs lightweight phrase matching and simple structural heuristics. In a detection skill, overstating capability can cause operators or automated systems to treat weak heuristics as authoritative attribution, producing false confidence, misclassification, and unsafe policy decisions.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal