Codex Auth Cleaner

• Outcome: Pass
• Latest audit: May 28, 2026 at 9:49 AM UTC
• Version:
  1.0.3

Overview

This skill is a disclosed CPA admin utility for cleaning Codex auth files, but it should be used carefully because it can disable and delete auth records.

Install only if you intend to grant CPA administrative access. Protect config.json because it may contain the CPA admin key, start with status before running check/delete/clean, and use monitor mode only when you are comfortable with repeated automated cleanup of Codex auth files that fail quota checks.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (3)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 85% confidence
• Finding: The skill exercises network access, reads environment variables, and writes persistent files, yet it declares no permissions. That mismatch can cause operators or calling agents to invoke it without understanding that it stores credentials locally and talks to an external management API, increasing the chance of over-privileged or unsafe execution.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 95% confidence
• Finding: The stated purpose focuses on cleaning invalid auth files, but the documented behavior also includes interactive setup, persistent credential storage, connectivity testing, monitoring loops, and notification configuration. This materially expands the trust boundary: the skill is not just a one-shot cleaner, it is a credential-handling, stateful, networked automation component that could expose admin keys or perform ongoing destructive actions if misused.

Vague Triggers

Medium

• Confidence: 78% confidence
• Finding: The activation phrases are broad for a skill that can disable and delete authentication files. Broad triggers increase the risk that an agent invokes a destructive auth-management workflow from vague user language, causing unintended cleanup actions against sensitive credentials.

VirusTotal

64/64 vendors flagged this skill as clean.