arize-dataset
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes datasets containing user-defined fields (e.g., question, answer, context) which are ingested into the agent context and could contain malicious instructions designed to influence agent behavior during evaluation or experimentation tasks.\n
- Ingestion points: Data enters the agent's context through
ax datasets export(file or stdout),ax datasets get(metadata and version lists), andax datasets list.\n - Boundary markers: The instructions do not specify the use of clear delimiters or instructions to ignore embedded commands when the agent processes the content of exported examples.\n
- Capability inventory: The agent can execute shell commands via the
axCLI, perform file system operations (reading and writing files for import/export), and perform network operations via the CLI's API interactions.\n - Sanitization: No explicit sanitization or validation of the content within the dataset examples is mentioned before the data is processed in downstream tasks like experiments or prompt optimization.\n- [EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the
arize-ax-clipackage from public registries using standard tools like pip, uv, and pipx. This is a legitimate installation procedure for the vendor's official command-line interface and does not present an unusual security risk.
Audit Metadata