The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external image assets which creates a potential surface for instructions embedded in those assets to influence the agent.
Ingestion points: The referenceAssetIds parameter used in the submit_image tool (found in SKILL.md).
Boundary markers: Absent. There are no specific delimiters or instructions provided to the model to ignore instructions that might be contained within the reference data.
Capability inventory: The skill has the ability to initiate image generation jobs (submit_image), monitor job status (track_progress), and manipulate project items (edit_item).
Sanitization: Absent. The instructions do not specify any validation or sanitization steps for the contents of the reference assets.
[SAFE]: The skill includes a reference to official documentation from a well-known service (OpenAI) for additional technical context regarding model parameters and best practices.

image-gen