ClawHub
Back to skill

Security audit

Codex Orchestrator

Security checks across malware telemetry and agentic risk

Overview

This skill matches its orchestration purpose, but it gives broad unattended control over Codex sessions with weak review gates around approvals and termination.

Install only if you intentionally want an agent to supervise autonomous Codex runs. Use it in a sandboxed repository or disposable branch, review logs before every submit action, avoid generic `y` or Enter approvals unless the prompt is clearly understood, track session IDs, and stop background sessions when work is complete.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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

High
Confidence
96% confidence
Finding
The skill explicitly instructs the operator to send 'y' or an empty Enter to an interactive prompt without requiring verification of what is being approved. In this context, the supervised Codex session may be asking for permission to perform sensitive actions, so blindly forwarding approval can authorize destructive commands, data exfiltration, dependency installation, or other unintended operations.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The kill guidance tells the operator to terminate a session when it appears to be looping or hallucinating, but it omits any warning that termination may interrupt in-flight edits, leave partial changes, or discard unsaved progress. In an orchestration skill, that can cause integrity and availability issues by corrupting task state or losing work without the user understanding the consequences.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.