ClawHub
Back to skill

Security audit

Agent of Empires

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward guide for managing coding-agent sessions, but it documents prompt-bypass and force-delete workflows without enough guardrails.

Review this skill before installing if you use aoe on real repositories. Avoid YOLO mode unless the agent is tightly sandboxed, and do not use force-delete worktree commands unless you have verified the exact target, checked for uncommitted work, and are comfortable losing that workspace.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Missing User Warnings

High
Confidence
98% confidence
Finding
The skill documents `aoe remove <id-or-title> --delete-worktree --force` as a routine command without warning that it can irreversibly delete a worktree and bypass confirmation. Because this skill manages coding workspaces and parallel worktrees, an operator or agent following the documentation could cause destructive data loss or remove active development environments without review.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill explicitly documents `aoe add ... -y -l` as a normal usage pattern while stating only that it will 'skip permission prompts', with no warning that this removes a key safety control for agent actions. In a skill whose purpose is managing autonomous coding agents, encouraging prompt bypass materially increases the chance of unreviewed file changes, command execution, or other unsafe actions by downstream agents.

Static analysis

No suspicious patterns detected.