codex
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
codex execcommand to delegate tasks to an external agent. This is the primary function of the skill and is documented for use in code review and architectural analysis. - [DATA_EXPOSURE]: The skill accesses local project files (e.g.,
src/auth/) to perform analysis. While it does not show signs of exfiltration, it operates within the user's project context. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8).
- Ingestion points: The skill reads user tasks and project files (e.g.,
src/auth/). - Boundary markers: None present; content is processed directly by the Codex agent.
- Capability inventory: Uses
codex execto execute tasks andcodex agentfor logic delegation. - Sanitization: No evidence of sanitization or filtering for embedded instructions within analyzed source code.
Audit Metadata