ClawHub
Back to skill

Security audit

Skill Compliance

Security checks across malware telemetry and agentic risk

Overview

This is a local compliance-checking skill whose executable behavior matches its stated purpose, with a plugin-loading design users should understand before extending it.

Install only if you are comfortable running a local Python scanner over the skill directories you choose. Do not add third-party files to the plugins directory unless you trust them, because plugin files execute as code during scans.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (2)

Context-Inappropriate Capability

Medium
Confidence
96% confidence
Finding
The checker dynamically discovers and executes Python plugins from a local plugins directory by modifying sys.path and calling each plugin's check() function. That turns a supposedly static/offline compliance scanner into an arbitrary code execution surface: anyone who can place or alter plugin code can run unrestricted Python during a scan, including filesystem access and subprocess/network activity via the plugin implementation.

Intent-Code Divergence

Medium
Confidence
88% confidence
Finding
The module-level description claims there are no network requests and only standard-library behavior, but the runtime plugin mechanism can import and execute arbitrary external code with unrestricted behavior. This mismatch is dangerous because users may trust the tool as offline/safe and run it in sensitive environments, while plugins can silently add network, process execution, or data exfiltration capabilities.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dynamic_code_execution

Dynamic code execution detected.

Critical
Code
suspicious.dynamic_code_execution
Location
plugins/__init__.py:46