ClawHub
Back to skill

Security audit

AI Safety Audit

Security checks for vulnerabilities and agentic risk

Overview

This is a non-executable AI safety audit checklist skill with some broad documentation language but no hidden commands, persistence, or credential handling.

Before using it, provide only the business and AI-system details needed for the audit, and avoid sharing secrets, credentials, raw customer data, or regulated personal information. Treat the external paid resource links as optional third-party links to vet separately.

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 (1)

Vague Triggers

Low
Confidence
89% confidence
Finding
The README invites users to 'Run this skill' for a complete audit outcome without describing clear trigger conditions, input boundaries, or when the skill should or should not be used. Broad invocation language can cause oversharing of sensitive business, compliance, or deployment details to the skill in contexts where a narrower scope would be safer.

Static analysis

No suspicious patterns detected.