The Agent Skills Directory

[PERSISTENCE]: The skill provides instructions for persisting the ARIZE_SPACE setting by modifying shell profiles (~/.zshrc, ~/.bashrc) or setting environment variables on Windows. While these are common persistence locations, they are used here for legitimate user configuration of the Arize workspace.
[COMMAND_EXECUTION]: The core functionality of the skill relies on executing ax CLI commands to create, read, update, and delete AI integrations.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions for installing the arize-ax-cli package from public registries using standard tools like pip, uv, and pipx.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill ingests data from the output of ax commands (e.g., ax ai-integrations list), which includes information retrieved from the remote Arize platform.
Boundary markers: There are no explicit delimiters or boundary markers specified to help the agent distinguish between tool-returned data and potentially malicious instructions that could be embedded in the platform data.
Capability inventory: The skill can execute various ax commands that modify remote state, such as creating or deleting integrations.
Sanitization: The skill does not implement explicit validation or sanitization of the CLI output before processing it in the agent context.

arize-ai-provider-integration