AI Agent Helper

• Outcome: Pass
• Latest audit: May 28, 2026, 2:00 AM
• Version:
  1.0.0

Overview

This is a small instruction-only skill for helping with AI agent prompts and setup, with no code or requested system access.

Install this if you want help drafting prompts or AI-agent workflows. Review any generated system prompts before using them with agents that can access private data, run tools, or take real-world actions.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The invocation examples are very broad and can cause the skill to activate for vague requests like 'help me write a prompt' or 'how to set up an AI agent' without clear boundaries. In an agent-assistance skill, overbroad triggering increases the chance of unsolicited prompt shaping or interference with unrelated tasks, which can change agent behavior in ways the user did not explicitly request.

Natural-Language Policy Violations

Medium

• Confidence: 79% confidence
• Finding: The description is written in Cantonese/Chinese and implies a language preference without stating that it adapts to the user's chosen language. This can cause unwanted language forcing or reduced usability, and in security-sensitive agent workflows it may obscure instructions or outputs if the user expects a different locale.

VirusTotal

64/64 vendors flagged this skill as clean.