codex-review
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands, specifically
codex exec reviewandgit, to analyze code and determine repository states. - [DATA_EXFILTRATION]: Repository source code and version history are transmitted to an external service (OpenAI Codex) for review. While this is the skill's primary function, it involves sending internal data to an external third-party service.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and evaluates repository content that could contain hidden malicious instructions.
- Ingestion points: Git diffs and repository files processed by the
codexcommand. - Boundary markers: Absent; no specific delimiters are used to separate code from potential instructions during the review process.
- Capability inventory: Shell execution (
codex,git) and filesystem access (/tmp/codex-review.txt). - Sanitization: No sanitization of input files or diff data is performed prior to analysis.
Audit Metadata