codex-review-branch

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The analyzed file is a metadata-only configuration for a code review variant. No malicious patterns, hardcoded credentials, obfuscation, or dangerous remote code execution patterns were found. All mentioned tools and resources belong to the expected environment for code analysis.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted data from feature branches. While this represents a surface for indirect prompt injection, it is an inherent characteristic of code review automation.
  • Ingestion points: Untrusted content is ingested from feature branches via Bash(git:*) operations.
  • Boundary markers: None are defined in this entry-point configuration.
  • Capability inventory: The skill utilizes Bash and Codex MCP tools to analyze and interact with code.
  • Sanitization: No explicit sanitization or filtering of the branch content is documented within this specific skill file.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 03:38 AM
Security Audit — agent-trust-hub — codex-review-branch