Security audit

Agent of Empires

Security checks across malware telemetry and agentic risk

Overview

The skill is coherent for managing AI coding-agent sessions, but it documents reduced-oversight and force-deletion workflows without enough safety guidance.

Review this skill before installing if you expect agents to operate only with explicit approvals. Avoid the YOLO example unless you are in a disposable or tightly sandboxed workspace, and do not use forceful worktree deletion until you have checked git state and confirmed the worktree contains nothing you need.

SkillSpector

By NVIDIA

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 (2)

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill explicitly advertises `-y`/YOLO mode as skipping permission prompts, but provides no warning about the loss of human approval gates. In a skill for managing AI coding agents, this can normalize launching agents with reduced oversight and increase the chance of unintended code changes, destructive commands, or data access without operator review.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The documented command `aoe remove <id-or-title> --delete-worktree --force` is destructive and shown without any warning that it may permanently delete worktree data. Because this skill is designed to be followed operationally by an agent or user, presenting forceful deletion as a normal example increases the risk of irreversible data loss.

VirusTotal

VirusTotal engine telemetry is currently stale for this artifact.

View on VirusTotal

Static analysis

No suspicious patterns detected.