ClawHub
Back to skill

Security audit

Operator Discipline

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only agent behavior guide with broad but disclosed scope and no executable code, hidden data access, or persistence mechanism.

Install this only if you want a broad behavioral discipline layer for agent configuration or prompt-writing work. Review the universal wording before embedding it into a system prompt, especially where local policy, accessibility needs, or user preferences should take precedence.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The skill advertises automatic activation for broad categories like agent configuration, system prompt design, and behavioral rule-setting, which are common meta-workflows. That can cause the skill to apply outside narrowly intended situations and silently override or bias other instructions, increasing the chance of unwanted behavioral changes across many sessions.

Natural-Language Policy Violations

Low
Confidence
83% confidence
Finding
The instruction to apply these rules in any session or system prompt regardless of persona, platform, or task domain is effectively universal and does not preserve room for user preference, local policy, or context-specific exceptions. In practice, this can create instruction conflicts, reduce operator control, and cause the agent to impose a rigid behavioral layer where it is inappropriate.

Static analysis

No suspicious patterns detected.