ClawHub
Back to skill

Security audit

抖音账号订阅追踪

Security checks across malware telemetry and agentic risk

Overview

The skill mostly does what it says, but it under-discloses persistent local storage and silent daily monitoring changes.

Review this before installing if you care about privacy or cleanup: queried Douyin IDs and the RedFox API key are used with redfox.hk, reports are written locally and may auto-open, and despite the documentation's no-local-storage claim the script can persist subscription and failure state in ~/.qoder. Confirm any daily automation before enabling it and know how to remove scheduled tasks and local files.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
78% confidence
Finding
The skill invokes capabilities including environment access, filesystem I/O, network requests, and shell execution, but does not declare any permissions. This weakens reviewability and consent because users and the platform cannot accurately assess what the skill can access before it runs. In this context, the skill also handles API keys and writes reports/files, so the undeclared capability surface is materially relevant rather than theoretical.

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
The skill claims that subscription state is kept only in automation command arguments and that it does not rely on local file storage, yet the analyzed behavior indicates persistent local storage of subscriptions and failure-tracking state. This is a serious trust and transparency issue because users may believe they are avoiding persistence while the skill actually retains monitored account identifiers and operational metadata on disk. In a subscription/monitoring skill, hidden persistence increases privacy, audit, and cleanup risks.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill instructs the agent to silently perform subscription changes and scheduled automation updates, while only showing the final result to the user. Silent creation or modification of persistent daily monitoring tasks reduces informed consent and can cause users to unknowingly enable ongoing surveillance, repeated API usage, or stored monitoring state. The risk is elevated here because the skill explicitly creates daily 9:00 automation jobs and may update them behind the scenes.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The script sends user-supplied Douyin account IDs and associated query activity to a third-party service (redfox.hk) without any explicit consent prompt, privacy notice, or service-boundary disclosure. In an agent context, users may reasonably assume lookups are local or first-party, so silent transmission can leak monitoring targets, interests, and operational data to an external provider.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.