ClawHub
Back to skill

Security audit

ai-content-creator-assistant

Security checks across malware telemetry and agentic risk

Overview

This paid content-writing skill is mostly clear and purpose-aligned, but it stores and prints the user's question locally as part of its payment order flow.

Install only if you are comfortable with the skill writing your full request to a local order file and printing it in command output. Do not include confidential business plans, personal data, unpublished source material, or sensitive credentials in the question unless you are comfortable with that local persistence.

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)

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The script documentation claims that only payment metadata is stored and explicitly says no created content or source materials are persisted, but the implementation stores the full user-supplied question in the order JSON. If users include sensitive prompts, business plans, personal data, or draft content, that data is written to disk under the user's home directory and may be exposed to other local processes, backups, or later collection.

Intent-Code Divergence

Medium
Confidence
98% confidence
Finding
The runtime notice tells users that no created content or source materials are stored, yet the script saves the user's question to the order file and also echoes it to stdout. This misrepresentation can cause users to disclose sensitive material under false assumptions, increasing privacy and data handling risk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.