orchestrating-codex
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill includes bash scripts (
scripts/codex-consult.shandscripts/codex-plan-review.sh) that execute thecodexCLI tool. These scripts use 'exec' to run the CLI with user-supplied arguments such as project directories and planning files. - [EXTERNAL_DOWNLOADS]: The documentation references the
@openai/codexNode.js package and instructs the user to install it via npm. This is a reference to a well-known service and aligns with the skill's primary purpose of providing a CLI integration. - [DATA_EXPOSURE]: The skill instructions mention accessing and managing agent definition files within the
~/.codex/agents/directory, which is the standard configuration location for the integrated Codex tool. - [PROMPT_INJECTION]: The instructions contain detailed guidelines for agent behavior and workflow orchestration. No patterns indicating safety bypasses, instruction overrides, or unauthorized system prompt extraction were detected.
Audit Metadata