ClawHub
Back to skill

Security audit

llm-behavior-distill

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent model-behavior helper with no evidence of hidden data access, persistence, credential use, or destructive behavior.

Installers should understand that this skill may shape the agent toward Chinese-language examples and automatic module routing, but the inspected package does not show hidden authority, data exfiltration, persistence, or destructive behavior.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Rogue AgentSelf-Modification, Session Persistence
  • 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)

Natural-Language Policy Violations

Medium
Confidence
86% confidence
Finding
The skill hard-codes Chinese as the default target/output language in the multi-language module and examples, which can override user preference and reduce transparency. This is not an exploit primitive by itself, but it can cause unintended behavior, miscommunication, or policy noncompliance in contexts where language choice must follow user intent or system settings.

Self-Modification

High
Category
Rogue Agent
Content
| `long_ctx` | Yi-1.5 | 长上下文处理 | 200K+文档无损分析 |
| `chunk_sum` | MiniMax | 线性注意力分块 | 无限长文档分层摘要 |
| `subagent` | Kimi K3 | 子Agent隔离上下文 | 并行任务独立执行 |
| `self_evolve` | Llama 4 | 社区微调模板 | 自我进化模板系统 |
| `compress` | Mistral | 高效压缩 | 50% token节省 |
| `guard` | Gemma 3 | 安全对齐 | 自检+护栏系统 |
| `small_reason` | Phi-4 | 小模型推理 | 轻量场景下深度思考 |
Confidence
80% confidence
Finding
self_evolve

VirusTotal

58/58 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.