ClawHub

OpenClaw Skill Advisor

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only advisor that recommends OpenClaw skills and installation commands, with no hidden code or automatic system access.

Safe to install as a lightweight advisor, but do not treat its recommendations as vetted approvals. Before running any suggested clawhub install command, review the recommended skill's publisher, files, permissions, scan results, and documentation, and verify any claimed ratings or download counts in ClawHub.

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 (3)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger phrases are broad, generic requests like "找技能" and "推荐技能" that can easily appear in ordinary conversation without the user explicitly intending to invoke this specific skill. This can cause unintended activation and increase the chance that the agent follows this skill's recommendation/install flow in contexts where the user only asked a general question.

Vague Triggers

Medium
Confidence
92% confidence
Finding
The statement that any listed phrase will activate the skill does not define boundaries, precedence, or disambiguation rules. In an agent environment with multiple skills, this ambiguity can lead to accidental routing, unexpected tool behavior, or the user being steered into installation advice without clear consent.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill provides terminal-based installation guidance and example install commands but does not clearly warn that installing third-party skills can change the local environment, add dependencies, or introduce untrusted code. In a recommendation skill, this is especially risky because users may treat the guidance as vetted and execute commands without reviewing the target package.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal