The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to echo the ARIZE_API_KEY environment variable to the terminal for verification, exposing sensitive credentials in the agent's output and session logs.
[COMMAND_EXECUTION]: The skill automates the modification of shell configuration files (e.g., ~/.zshrc, ~/.bashrc) and uses PowerShell to set persistent user-level environment variables to save credentials. These are persistence mechanisms that modify the user's system environment.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the arize-ax-cli package via pip, uv, or pipx. While this is a vendor-provided tool, it involves downloading external code from a registry at runtime.
[PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection by ingesting untrusted data through commands like ax datasets create and ax datasets append from files or JSON strings.
Ingestion points: ax datasets create --file and ax datasets append (via --json or --file).
Boundary markers: Absent; no instructions are provided to delimit or ignore instructions embedded in the dataset content.
Capability inventory: Includes shell command execution, software installation, and system environment modification.
Sanitization: Absent; the skill does not include steps to validate or sanitize input data before processing.

arize-dataset