codex-code-review
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local CLI tools including 'codex', 'git', and 'backlog' to facilitate automated code reviews, manage git history (via amends), and track remediation progress in the project backlog.- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface (Category 8) by ingesting untrusted data from external review files and using that data to drive agent actions.
- Ingestion points: The agent reads finding descriptions and 'Suggested fix' instructions from markdown files located in '.agent/reviews/'.
- Boundary markers: None. The skill does not implement delimiters or warnings to ignore malicious instructions embedded in the review output.
- Capability inventory: The skill can modify source code, execute git commands, and create backlog tasks.
- Sanitization: Absent. The skill parses findings using shell scripts and regex without verifying the integrity or safety of the suggested fixes before implementation.
Audit Metadata