run-codex-exec

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute local scripts, including a Node.js dispatcher (node skills/use-codex/skills/use-codex/scripts/use-codex.mjs) and a detached Bash runner (bash scripts/run-fleet.sh). These tools orchestrate parallel coding tasks and manage the git lifecycle across multiple worktrees.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing user-controlled input from a tasks.json file which is used to drive automated coding agents.\n
  • Ingestion points: User-provided tasks.json file containing discrete coding tasks (referenced in 'How to route' section).\n
  • Boundary markers: The shim does not implement explicit boundary markers or delimiters for ingested data, relying instead on downstream contracts within the 'use-codex' skill.\n
  • Capability inventory: The skill possesses the ability to execute shell scripts, Node.js applications, manage git worktrees, and perform automated commits.\n
  • Sanitization: No specific sanitization or validation of input file content is described in the shim instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 02:36 PM
Security Audit — agent-trust-hub — run-codex-exec