ClawHub
Back to skill

Security audit

面试助手

Security checks across malware telemetry and agentic risk

Overview

This interview-coach skill does what it says, but users should know it may install Python parsing libraries and may upload image or scanned-PDF contents to an OCR service.

Install only if you are comfortable with the skill installing Python packages at first use. For sensitive resumes or JDs, prefer text, Word, or text-based PDF input so content stays local; image and scanned-PDF OCR sends the document and OCR token to the configured PaddleOCR endpoint.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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 (3)

subprocess module call

Medium
Category
Dangerous Code Execution
Content
import subprocess
    print(f"[安装依赖] {pkg} ...", file=sys.stderr)
    try:
        subprocess.check_call(
            ["uv", "pip", "install", pkg, "--system", "-q"],
            stderr=subprocess.DEVNULL,
        )
Confidence
94% confidence
Finding
The script performs runtime package installation via an external package manager, which executes code fetched from package repositories during normal skill operation. In an agent environment this expands the trust boundary, enables supply-chain compromise or unintended environment mutation, and is not necessary for safe parsing if dependencies are predeclared.

subprocess module call

Medium
Category
Dangerous Code Execution
Content
except FileNotFoundError:
        print("[安装依赖] uv 未找到，改用 pip ...", file=sys.stderr)
        try:
            subprocess.check_call(
                [sys.executable, "-m", "pip", "install", pkg, "-q"],
                stderr=subprocess.DEVNULL,
            )
Confidence
94% confidence
Finding
The fallback call to pip also installs packages at runtime, again causing arbitrary third-party code execution and modifying the host Python environment. Because this occurs automatically on missing imports, a user-triggered parse path can lead to network access and package installation without an explicit administrative approval step.

Context-Inappropriate Capability

Medium
Confidence
96% confidence
Finding
Automatic dependency installation is unjustified for a document parsing helper and creates a clear supply-chain and environment-integrity risk. Since this skill processes untrusted user files, combining untrusted input handling with self-modifying runtime behavior makes the operational context more dangerous, not less.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.