The Agent Skills Directory

[SAFE]: The skill enforces security best practices for credential handling by explicitly forbidding the agent from reading secrets from files (like .env) or asking users to paste API keys into the chat. It instructs the agent to use environment variables for sensitive values in CLI commands.\n- [SAFE]: External dependencies are restricted to the vendor's official CLI tool (arize-ax-cli), ensuring that the toolchain is authentic and provided by the skill author.\n- [SAFE]: Local file system access is highly restricted and limited to reading specific, non-sensitive configuration values (ARIZE_SPACE_ID) if required for session context.\n- [SAFE]: Persistence mechanisms are limited to the standard configuration of the ax CLI tool in the user's home directory (~/.arize/config.toml), which is necessary for the intended functionality of managing profiles.

arize-ai-provider-integration