Skills
skills/yeachan-heo/oh-my-codex/omx-setup/Gen Agent Trust Hub

omx-setup

Pass

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from AGENTS.md, creating a surface for indirect prompt injection. \n
  • Ingestion points: The documented setup process reads and merges content from AGENTS.md in the project root. \n
  • Boundary markers: Employs explicit delimiters (<!-- OMX:AGENTS:START --> and <!-- OMX:AGENTS:END -->) to distinguish managed content from user-authored sections. \n
  • Capability inventory: The omx setup command performs file system modifications to core configuration and hook files. \n
  • Sanitization: No explicit sanitization or filtering of the user-authored instructions within AGENTS.md is described. \n- [COMMAND_EXECUTION]: The skill instructs the agent on executing setup procedures that modify local environment and configuration files. \n
  • File access: Modifies ~/.codex/config.toml and ~/.codex/hooks.json to configure tool behavior and MCP servers. \n
  • Environment management: Handles the configuration of CODEX_HOME and tool-specific state directories (.omx).
Audit Metadata
Risk Level
SAFE
Analyzed
May 6, 2026, 08:06 AM