The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently encourages the use of the --dangerously-bypass-approvals-and-sandbox flag. This disables essential safety mechanisms and allows the agent to execute code without user oversight or isolation.
[DATA_EXFILTRATION]: Instructions demonstrate how to read and extract authentication tokens and API keys using shell commands like pass show, security find-generic-password, and cat ~/.codex/config.toml. While used for authentication, these patterns involve accessing sensitive credentials.
[CREDENTIALS_UNSAFE]: The documentation includes multiple examples of hardcoding API keys in configuration files and environment variables, such as export OPENAI_API_KEY="sk-your-api-key-here" and api_key = "sk-your-api-key-here" in ~/.codex/config.toml.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @openai/codex package globally via npm install -g @openai/codex. While OpenAI is a well-known service, the skill also suggests installing the pass utility from general package registries (apt-get, brew).

codex-auth