ClawHub
Back to skill

Security audit

qq-zone-photo-free

Security checks across malware telemetry and agentic risk

Overview

The skill is not clearly malicious, but it handles full account cookies and asks for command/write access while leaving key implementation and outbound-callback behavior under-specified.

Review this before installing if the QQ Zone account contains private or important data. Use a dedicated cookie file with restrictive permissions, avoid supplying a callback URL unless you know exactly what it will receive, and do not provide cookies or API keys beyond what is needed for login, listing, and photo browsing.

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
93% confidence
Finding
The skill advertises a limited free feature set focused on browsing, but still requests both exec and write capabilities, which are broader than necessary for the documented scope. Excess permissions increase the blast radius if the skill instructions, underlying script, or future edits cause unintended file modification or command execution, especially because cookies and other local credentials are involved.

Intent-Code Divergence

Medium
Confidence
84% confidence
Finding
The documentation assures users that credentials remain local and are not uploaded, yet the skill uses non-official APIs and also accepts a callback URL, creating a mismatch between the trust claim and actual data-flow possibilities. That inconsistency can mislead users into exposing sensitive cookies or account data under weaker assumptions than the system actually enforces.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.