The Agent Skills Directory

[SAFE]: The skill serves as a prompt engineering and workflow guide for Grok Imagine video generation. It does not contain executable code, hardcoded credentials, or obfuscated content.
[INDIRECT_PROMPT_INJECTION]: The skill acknowledges the risk of malicious instructions being embedded in user-supplied reference images, videos, or prompt documents. It includes specific defensive instructions for the agent to treat these as untrusted data rather than commands, which is a security best practice for media-processing skills.
[EXTERNAL_DOWNLOADS]: The skill references official xAI developer documentation URLs for verifying current model capabilities and pricing. These references are to well-known, trusted domains related to the skill's primary purpose.
[PROMPT_INJECTION]: The skill includes robust safety guidelines that refuse requests to generate deceptive synthetic media, impersonations, or fake-event footage, even if the user attempts to bypass these rules with 'parody' or 'satire' labels.

grok-imagine