Cool Agent Tools

• Outcome: Review
• Latest audit: May 29, 2026 at 9:34 PM UTC
• Version:
  1.0.0

Overview

This is a disclosed operations command cheat sheet, but it gives an agent broad permission to run disruptive shell commands without clear safeguards.

Install only if you want an agent to help with system administration tasks. Before using it, require explicit approval for delete, kill, background, Docker-changing, system-path, or outbound-network commands, and ask the agent to preview affected files, processes, and targets before making changes.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (3)

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The skill says it needs no special trigger and can be called directly during general system analysis or operations tasks, which makes activation scope extremely broad. In an agent context, this increases the chance that potentially dangerous shell commands are invoked without explicit user confirmation or narrow task scoping, especially because the same skill also documents destructive operations like deletion and process termination.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This section documents forceful and destructive commands such as pkill, kill -9, log deletion, and broad process termination without warnings, safeguards, or confirmation requirements. In an agent skill, presenting these as normal utility actions can lead to service disruption, data loss, or accidental termination of unrelated workloads if the agent applies them too broadly or on ambiguous requests.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The network diagnostics section encourages commands like curl, ping, and nslookup against arbitrary hosts without warning about external connectivity, metadata leakage, or policy restrictions. In an agent setting, this can cause unintended outbound requests to attacker-controlled endpoints, leak internal network information, or violate environment network controls.

VirusTotal

64/64 vendors flagged this skill as clean.