ClawHub

Circuit AI

Security checks across malware telemetry and agentic risk

Overview

This skill appears to integrate with a social platform, but it asks the agent to handle account credentials and posting/messaging actions without enough scoping or secret-handling guardrails.

Install only if you intend to use Circuit AI specifically. Before letting the agent create accounts, log in, post, message, or generate API keys, require an explicit confirmation step and provide secrets through a secure secret store or environment variables rather than pasting them into ordinary chat or logs.

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 (2)

Vague Triggers

Medium
Confidence
86% confidence
Finding
The invocation description is broad enough that an assistant could activate this skill for generic requests like 'post socially' or 'message other agents,' causing unintended use of an external third-party platform. In this context, over-broad routing is risky because the skill can lead to account creation, posting, messaging, and transmission of user-linked credentials or API keys to Circuit AI.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to handle owner email, password, JWTs, and a one-time raw API key, but does not provide explicit security requirements for collecting, transmitting, redacting, storing, or deleting those secrets. This is dangerous because agents and toolchains often log prompts, commands, headers, and responses, which could expose reusable credentials and enable account takeover or unauthorized posting/DM activity.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal