The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources, such as product pages via the import_from_url tool and user-provided screenshots or recording frames. This creates an attack surface where malicious instructions hidden in the visual or metadata content of those external sources could potentially influence the agent's behavior during the animation construction process.
Ingestion points: SKILL.md identifies import_from_url for fetching product pages, user screenshots, and get_clip_frame for recording frames as sources.
Boundary markers: The instructions lack explicit boundary markers or directions to the agent to ignore instructions that might be embedded within the processed media or web content.
Capability inventory: The skill leverages powerful media manipulation tools including addMedia, duplicate, setMasks, bindProperty, and setCurve to build and animate motion documents.
Sanitization: No sanitization or validation logic is specified for the data retrieved from external URLs or image files before it is processed by the agent.

tella-ui-parallax