ClawHub
Back to skill

Security audit

Drop Pick

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a product-research helper, but it also includes store-import and listing-event commands that can change commerce accounts without a clear approval boundary.

Install only if you want an agent to research Alibaba, AliExpress, and Amazon product opportunities using your Alibaba credentials. Treat the import and listing-event commands as live commerce actions: require explicit approval before running them, use scoped credentials where possible, and review the output path before allowing report files to be written.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill’s declared purpose is research/report generation, but it also includes operational actions that can modify external commerce systems by importing products and sending marketplace listing events. This expands the capability from analysis into side-effecting business actions, creating a confused-deputy risk where a user invoking research could unintentionally trigger store or marketplace changes.

Context-Inappropriate Capability

Medium
Confidence
97% confidence
Finding
A channel import capability is not necessary for product research and materially increases the blast radius of the skill by enabling direct publication or synchronization to storefronts. In context, this unjustified privilege violates least privilege and could be abused or accidentally invoked to push unwanted products into a merchant’s sales channels.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The invocation text contains very broad trigger phrases such as general product research and ‘what should I sell,’ which can cause the skill to activate for ordinary planning conversations. Because the skill includes file-writing and references to side-effecting commerce operations, overbroad routing increases the chance of unnecessary tool use and unintended actions.

Missing User Warnings

Low
Confidence
84% confidence
Finding
The skill instructs writing a report into the working directory without warning the user or obtaining consent for filesystem modification. While the write target appears limited and not inherently destructive, silent file creation can still surprise users, overwrite expected outputs, or be chained with other behaviors in automation contexts.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.