The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the ax command-line utility to interact with the Arize platform. These commands (e.g., ax datasets list, ax datasets export) are the core functionality of the skill and are used in a standard manner.
[EXTERNAL_DOWNLOADS]: The documentation references the arize-ax-cli package, which can be installed from the official Python Package Index (PyPI). This is a legitimate dependency required for the skill to function.
[DATA_EXFILTRATION]: The skill handles sensitive data such as API keys and space IDs. It demonstrates a high level of security awareness by instructing the agent never to read .env files or ask users for their raw keys in the chat, instead recommending standard environment variable practices.
[PROMPT_INJECTION]: The skill processes external data retrieved from the Arize platform (e.g., dataset examples). While this constitutes a potential indirect prompt injection surface, it is inherent to the skill's primary purpose of data management.
Ingestion points: ax datasets export and ax datasets get in SKILL.md.
Boundary markers: Absent; the skill does not specify markers to delimit retrieved data from instructions.
Capability inventory: The skill can execute ax and jq commands.
Sanitization: No explicit sanitization or filtering of platform data is mentioned.

arize-dataset