ClawHub

Quantum Lab

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed local runner for a specific quantum_lab Python project, with normal command-execution risk but no evidence of hidden or malicious behavior.

Install only if you trust the local quantum_lab repository, notebooks, and ~/.venvs/qiskit environment. Use the listed qexec commands deliberately, review requirements.txt before package installs, and avoid treating Telegram/OpenClaw requests as permission to run arbitrary local commands unless you trust the requester.

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)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding
The skill is presented as a helper for running specific quantum_lab scripts, but its interface and guidance allow users to invoke wrapper scripts with arbitrary arguments, effectively enabling general command execution within the repo directory and activated virtual environment. In an agent setting, this broad execution surface is dangerous because untrusted prompts could steer the agent to run unintended commands, install packages, or execute attacker-chosen code under the user's account.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal