ClawHub

AgentGuard Tech

Security checks across malware telemetry and agentic risk

Overview

This is a plausible security integration, but it would install an unpinned third-party SDK, create or use an external account, store an API key locally, and give that SDK broad control over all agent tool calls.

Review before installing. Use it only if you trust the AgentGuard provider and npm package, understand what tool names/arguments may leave your machine, can protect the stored API key, and have a tested rollback plan for restoring original tool behavior.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger phrases are broad and action-oriented, making accidental or overly permissive invocation more likely. In this skill's context, invocation can install an npm package, create an external account, and modify local configuration, so ambiguous triggers increase the chance of unintended side effects from casual user phrasing.

Missing User Warnings

High
Confidence
97% confidence
Finding
The skill does not clearly disclose that running it may install third-party code, create an external AgentGuard account, send data over the network, and store an API key locally. Missing this consent boundary is dangerous because users may invoke a 'security' skill expecting passive hardening, not external dependency installation and secret/config persistence.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The usage examples normalize one-shot invocation without warning that an account may be created automatically and that credentials/settings may be written locally. This increases the risk of users unknowingly provisioning external services and persisting secrets, especially because the examples are likely to be copied verbatim.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal