Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward CLI config editor, but it can persistently change security-sensitive Codex settings without built-in confirmation or rollback guidance.
Install only if you want an agent to help edit Codex CLI configuration. Before allowing writes, review the exact diff and avoid enabling Run Everything, broad tool allowlists, disabled sandboxing, or unrestricted network access unless you understand the persistent effect across future CLI sessions.
The skill description is broad enough to match many routine requests about CLI behavior, which can cause the skill to activate in situations where the user did not explicitly ask to edit configuration. Because the skill is capable of changing persistent settings in a sensitive config file, overbroad invocation increases the chance of unintended or surprise state changes.
The skill directly instructs the agent to read, modify, and write back a user configuration file without any warning, confirmation, backup, or discussion of consequences. Since this file controls permissions, sandboxing, networking, and approval behavior, silent modification could weaken protections or disrupt the user's environment persistently across sessions.
The documentation explicitly describes a mode that auto-approves all tool calls, equivalent to force/yolo behavior, without coupling it to strong warnings or restrictive workflow guidance. In the context of a skill that edits persistent CLI configuration, this can normalize or facilitate disabling an important safety control, increasing the risk of unauthorized or unsafe tool execution in future sessions.
### `approvalMode` (optional)
Controls tool approval behavior:
- `"allowlist"` (default) — require approval for tools not in the allow list
- Run Everything — auto-approve all tool calls (same as `--force` / `--yolo`; set approval mode to the non-allowlist value)
### `sandbox` (optional)
Sandbox execution environment settings:
No suspicious patterns detected.