codex-review
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Uses the
codexCLI and shell utilities likemktempto manage review artifacts and execute the analysis process. - [DATA_EXFILTRATION]: Transmits local repository content, including uncommitted changes and branch diffs, to OpenAI's infrastructure for the purpose of code review.
- [PROMPT_INJECTION]: Presents an attack surface for indirect prompt injection (Category 8) when analyzing untrusted code changes.
- Ingestion points: The skill reads uncommitted changes, git branch diffs, and implementation files as described in the Workflow section of
SKILL.md. - Boundary markers: No explicit delimiters or instructions are used to isolate the code being reviewed from the agent's instructions.
- Capability inventory: The skill executes shell commands via the
codexCLI and creates temporary files. - Sanitization: There is no evidence of input validation or content escaping for the code data processed by the tool.
Audit Metadata