run-codex-review

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill specifies shell commands to be executed for various routing scenarios, such as running node on local scripts (e.g., node skills/orchestrate-codex/skills/orchestrate-codex/scripts/orchestrate-codex.mjs). These commands are intended to be executed in the user's local environment.- [COMMAND_EXECUTION]: The skill performs prerequisite checks using CLI tools including codex --version, git rev-parse, and codex login status to verify the environment before routing the user to a canonical skill.- [PROMPT_INJECTION]: The skill ingests untrusted user input in the form of branch names and branch lists which are then interpolated into shell command templates. This represents an indirect prompt injection surface.
  • Ingestion points: User-provided branch names or lists from branches.txt as described in SKILL.md.
  • Boundary markers: None are present in the provided command templates to delimit the interpolated input.
  • Capability inventory: The skill facilitates the execution of local node scripts with access to the file system and project environment.
  • Sanitization: No explicit sanitization, validation, or escaping of the user-provided branch strings is defined within this routing shim.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 02:36 PM
Security Audit — agent-trust-hub — run-codex-review