Security audit

Codex

Security checks for vulnerabilities and agentic risk

Overview

This Codex integration plugin has powerful but disclosed session and runtime controls, and I found no hidden or purpose-mismatched behavior.

Install only if you want OpenClaw to integrate deeply with Codex. Review settings for user-home sharing, native session discovery, supervision raw transcript access, write controls, native plugin exposure, and execution/sandbox policy before enabling them.

SkillSpector

SkillSpector was not run because this plugin release contains no bundled skills.

Static analysis

Detected: suspicious.dangerous_exec, suspicious.exposed_secret_literal

Shell command execution detected (child_process).

Critical

Code: suspicious.dangerous_exec
Location: dist/.setup/dynamic-tools-D16lOp67.mjs:2082
Evidence: child = spawn(command, args, {

Shell command execution detected (child_process).

Critical

Code: suspicious.dangerous_exec
Location: dist/.setup/transport-stdio-Cmx94Oiy.mjs:132
Evidence: const inspector = execFile(procfs ? process.execPath : "ps", procfs ? [

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: dist/.setup/native-auth-BLZGqFjw.mjs:43
Evidence: apiKey: "[REDACTED]",