ClawHub
Back to skill

Security audit

Brainstorming

Security checks across malware telemetry and agentic risk

Overview

This is a strict design-planning skill that can slow workflows and create a committed spec, but its behavior is disclosed and aligned with its purpose.

Install this if you want a strict design-before-code workflow. Expect it to read repository context, ask for design approval before implementation, create a spec file, and make a git commit; decline or override those steps when you only want a quick edit or do not want repository changes.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • 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
Findings (2)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The skill description uses mandatory, highly broad trigger language ('You MUST use this before any creative work') that can force invocation for a very wide range of normal development activity. In an agent environment, this can create workflow hijacking or denial-of-service style friction by routing many unrelated tasks through this skill before any other action can occur.

Session Persistence

Medium
Category
Rogue Agent
Content
## Checklist

You MUST create a task for each of these items and complete them in order:

1. **Explore project context** — check files, docs, recent commits
2. **Offer the visual companion just-in-time** — NOT upfront. The first time a question would genuinely be clearer shown than described, offer it then (its own message); on approval its browser tab opens for you. If no visual question ever arises, never offer it. See the Visual Companion section below.
Confidence
77% confidence
Finding
The instruction to create and complete a task for each checklist item in order introduces persistent workflow state and can compel the agent to perform additional actions such as checking files, commits, writing docs, and invoking follow-on skills. In a skill context, this increases the chance of unnecessary stateful behavior, unintended tool use, and user-request expansion beyond the original task.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.