ClawHub
Back to skill

Security audit

ai-agent-helper

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly an AI-agent design guide, but it requests command execution and file-writing authority without tight upfront limits.

Review this before installing if you only wanted prompt-design advice. Use it in a least-privileged workspace, require explicit approval for any command execution, file writes, API calls, callbacks, or credential use, and do not run it with admin privileges unless you have a specific, trusted task that requires it.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The skill is presented as an AI agent design/optimization helper, but later documentation expands its role to operational capabilities such as file handling, API integration, and command execution. This mismatch can cause users or orchestrators to grant broader trust and permissions than intended, increasing the chance of unsafe invocation or privilege misuse.

Vague Triggers

Medium
Confidence
88% confidence
Finding
An overly broad invocation description makes it easier for the skill to be applied to ambiguous or high-impact tasks without clear boundaries. In a skill that advertises agent-loop design and has access to exec/write-capable tools, weak trigger constraints increase the risk of the agent drifting from advisory use into unsafe operational behavior.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill advertises file writing, API access, and command execution capabilities without prominent upfront warning about their consequences. Because the skill also lists powerful tools like exec and write, users may invoke it expecting harmless design guidance while unintentionally enabling actions that can modify files, run commands, or exfiltrate data.

Static analysis

No suspicious patterns detected.