The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill acts as an interface for processing untrusted user input and external assets which are then forwarded to AI models for processing.
Ingestion points: Untrusted data enters the context via the prompt and referenceImages parameters in the model_run tool, as described in the worked example and documentation.
Boundary markers: The skill instructions do not specify the use of delimiters or 'ignore' commands to prevent the model from following instructions that might be embedded in user-provided prompts or descriptions.
Capability inventory: The skill leverages tools such as model_run, upload_asset, and asset_download to perform image generation and file operations.
Sanitization: No sanitization, escaping, or validation of user-provided prompt strings or reference image content is described.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install related tools using the command npx skills add scenario-labs/skills. This targets the official repository of the skill author and is a routine operation within this platform's ecosystem.

scenario-gemini-image