The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references the installation of additional components using the command npx skills add scenario-labs/skills --skill <name>. This points to the official repository managed by the skill's author to extend functionality.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that processes user-supplied data, creating a surface for potential indirect prompt injection.
Ingestion points: Untrusted data is processed via the prompt, image, and video parameters within the model_run tool as described in SKILL.md.
Boundary markers: There are no explicit instructions in SKILL.md for using delimiters or negative constraints to prevent the agent from following instructions embedded in input data.
Capability inventory: The SKILL.md file outlines the use of tools including model_run, upload_asset, and search to interact with generative models.
Sanitization: The skill instructions do not specify any sanitization or validation steps for the inputs before they are passed to the underlying tools.

scenario-runway