ClawHub

Liuliu Proactive Agent

Security checks across malware telemetry and agentic risk

Overview

This skill is a Review item because it encourages an agent to monitor sensitive information and change files or code without clear user approval.

Install only if you want a more autonomous agent and are prepared to set strict rules. Require explicit approval before calendar or email monitoring, memory edits, file cleanup, document changes, code commits, or any action outside named accounts, paths, and time windows.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Confidence
94% confidence
Finding
The skill description uses very broad trigger terms like '主动', '积极', '提醒', 'proactive', and 'take initiative', which can match many ordinary user requests and cause the skill to activate outside its intended scope. Because this skill encourages unsolicited actions and proactive monitoring, accidental activation increases the chance the agent will overreach, act without explicit consent, or surface information the user did not request.

Missing User Warnings

High
Confidence
97% confidence
Finding
The skill explicitly says the agent can perform impactful actions without asking, including organizing memory files, updating documents, committing code changes, deleting temporary files, and checking project status. These are state-changing operations that can alter user data, codebases, or local environments; combined with the skill's emphasis on initiative, this creates a real risk of unauthorized modification, destructive cleanup, or unintended commits.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal