grok-media
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, specifically public image URLs and user-supplied prompts for media generation. This creates an attack surface where malicious content embedded in source images or prompts could potentially influence the agent's behavior.
- Ingestion points: SKILL.md describes using public HTTPS URLs and base64 data URIs as inputs for image and video generation tasks.
- Boundary markers: There are no boundary markers or instructions to ignore embedded commands within the processed data, which is a common practice for mitigating injection.
- Capability inventory: The skill facilitates network operations to the xAI API (https://api.x.ai/v1) using an API key.
- Sanitization: The documentation does not specify any sanitization or validation requirements for the external URLs or the data retrieved from them before they are passed to the model.
Audit Metadata