ClawHub
Back to skill

Security audit

Pressure Test

Security checks for vulnerabilities and agentic risk

Overview

This skill is a planning aid that asks an agent to pressure-test ideas and, when explicitly handed off, make reversible decisions with a visible record.

Install this if you want a structured pressure-testing workflow. Be aware that its sous mode lets the agent make reversible planning calls when you hand off work, so use clear instructions about what decisions it may make and review any saved decision transcript afterward.

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The skill defines sous-mode activation using broad, natural-language cues such as 'figure it out', 'I'm going AFK', or a user 'goes quiet after authorizing autonomous work'. These triggers are ambiguous and can cause the agent to shift from interactive clarification to autonomous decision-making without clear, explicit consent, increasing the chance of unintended actions or fabricated assumptions being treated as user-approved direction.

Static analysis

No suspicious patterns detected.