ClawHub

中文多智能体辩论

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Chinese debate-assistant skill with no code, credentials, external access, or persistence requirements.

Install if you want a Chinese-language structured debate workflow. Expect it to steer qualifying requests into a formal multi-round debate; ask explicitly for normal discussion or another language if that is not what you want.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation condition is broad enough to trigger on generic requests to debate, test arguments, or prepare a discussion, which can unexpectedly override the user’s preferred interaction mode. This creates scope-control risk: the agent may launch a long multi-agent workflow, generate unnecessary content, or steer benign conversation into a structured debate without clear consent.

Natural-Language Policy Violations

Medium
Confidence
82% confidence
Finding
The skill is explicitly Chinese-only and does not provide a language negotiation step, which can cause the agent to ignore user language preferences or produce outputs the user cannot verify. In security-sensitive contexts, forced language can reduce transparency and user oversight, increasing the chance of misunderstandings or unnoticed harmful content.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal