ClawHub
Back to skill

Security audit

Agent Debate

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward agent-debate workflow that creates local debate notes and does not show hidden, destructive, or data-exfiltrating behavior.

Before installing, expect the skill to create local markdown files in a plans/debate-{topic}/ folder. Use a non-sensitive topic name and review the generated debate notes before committing or sharing them.

SkillSpector

Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Missing User Warnings

Low
Confidence
93% confidence
Finding
The skill explicitly instructs agents to create and modify multiple files under `plans/debate-{topic}/` without warning the user that running the skill will write artifacts to the filesystem. This is not overtly malicious, but it can cause unexpected workspace changes, clutter, accidental overwrites, or disclosure of sensitive debate content into tracked directories if users are unaware.

Static analysis

No suspicious patterns detected.