codex-knowledge-patch

Pass

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill documentation details the use of the codex command-line utility for executing scripts, applying patches, and managing background terminals. It specifically mentions the ! syntax for queuing shell commands within the TUI environment.\n- [DATA_EXFILTRATION]: The instructions cover the configuration of remote WebSocket connections for TUI and MCP servers, including authentication via capability tokens and HS256 JWTs. These are described as standard features for remote workspace management.\n- [PROMPT_INJECTION]: The skill provides guidance on setting security boundaries using sandbox presets and approval policies. It documents advanced configuration flags such as --yolo for bypassing standard safety prompts, positioning them as expert features for isolated environments.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 20, 2026, 11:44 PM
Security Audit — agent-trust-hub — codex-knowledge-patch