ClawHub
Back to skill

Security audit

Inversion Strategist

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only strategy skill that gives inversion-style planning advice, with no code execution or persistent access, but users should be careful with its broad triggers and biometric example.

Install only if you want an inversion-based planning prompt. Be aware it may trigger from broad phrases like "flip it" or "what NOT to do," and do not treat the childcare face-recognition example as implementation guidance without privacy, consent, security, retention, and legal review.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

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 (3)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger list includes broad phrases such as "flip it," "opposite approach," and "what NOT to do," which can appear in many ordinary conversations unrelated to deliberate invocation of this skill. That raises the chance of accidental activation, causing the agent to switch into an inversion framing unexpectedly and potentially derail user intent or suppress more appropriate skills.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
Referencing face recognition only as a reliability issue ('fails often') frames the technology as an operational feature rather than a high-risk biometric processing activity. In this skill's business-advice context, that omission can encourage deployment without considering consent, child privacy, bias, security, and legal restrictions tied to processing minors' biometric data.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The content instructs operators to ensure all children's faces are enrolled before launch, which normalizes biometric collection for minors without any mention of parental consent, lawful basis, retention limits, or data protection safeguards. In a childcare context, omission of these privacy prerequisites can lead users to implement facial recognition in a way that violates privacy laws and exposes sensitive biometric data.

Static analysis

No suspicious patterns detected.