codex-plusplus-tweak-system

Fail

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: CRITICALREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill explicitly instructs users to download and execute code from an external source via high-risk shell pipes. This includes curl -fsSL https://raw.githubusercontent.com/b-nnett/codex-plusplus/main/install.sh | bash for Unix-based systems and irm https://raw.githubusercontent.com/b-nnett/codex-plusplus/main/install.ps1 | iex for Windows.
  • [COMMAND_EXECUTION]: The skill facilitates extensive interaction with the local shell to install, repair, and update the tweak system, including commands like codexplusplus install, codexplusplus update, and recursive directory deletion using rm -rf for troubleshooting.
  • [PRIVILEGE_ESCALATION]: The skill includes instructions to bypass macOS Gatekeeper security protections using the xattr -cr /Applications/Codex.app command and describes patching the app.asar binary, which compromises the integrity of the installed application.
  • [DYNAMIC_EXECUTION]: The skill is centered around a runtime that patches a desktop application to discover and execute ESM modules (tweaks) with full Electron renderer privileges. This allows for the runtime injection and execution of code not originally packaged with the application.
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface where instructions processed from external 'tweak' manifests or scripts could influence agent behavior or lead to malicious outcomes.
  • Ingestion points: Tweak directories such as ~/Library/Application Support/codex-plusplus/tweaks/ where the agent reads manifests and code.
  • Boundary markers: None; the skill does not provide clear delimiters or warnings for the agent to ignore potentially malicious instructions inside third-party tweaks.
  • Capability inventory: Binary patching, shell command execution, file system modification, and execution of arbitrary JavaScript/TypeScript code.
  • Sanitization: There is no documented validation or sanitization process for the contents of the tweaks before they are executed in the Electron environment.
Recommendations
  • HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/b-nnett/codex-plusplus/main/install.sh - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 16, 2026, 03:57 AM