ClawHub
Back to skill

Security audit

Python Dependency Candidate Fetcher 2

Security checks across malware telemetry and agentic risk

Overview

This skill is a local Python package recommendation helper with a noticeable bias toward Polars, but it does not show hidden execution, data theft, persistence, or destructive behavior.

Before installing, understand that this skill may steer tabular Python package choices toward Polars even when other candidates are plausible. It appears safe from a security perspective, but users who need neutral package selection should review or adjust the selection guidance.

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 (1)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The skill presents itself as a neutral selector over a fixed candidate list, but the embedded guidance systematically steers outcomes toward a specific package, especially polars, regardless of the caller’s intended neutrality. This creates instruction-level bias that can subvert upstream ranking logic, reduce reliability, and cause systematically incorrect or self-serving package recommendations when multiple candidates are plausible.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.