The Agent Skills Directory

[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it is designed to ingest and process data from external websites.
Ingestion points: Step 0 of the SKILL.md workflow instructs the agent to immediately fetch user-supplied URLs using WebFetch or a browser tool.
Boundary markers: The instructions do not specify any delimiters or safety guidelines to differentiate between the agent's logic and the potentially adversarial content found on the external websites.
Capability inventory: The agent uses the fetched information to generate a creative brief and subsequently a full HTML/JavaScript composition.
Sanitization: There are no sanitization or validation steps described to filter out malicious instructions that might be embedded in the target website's text or metadata.
[COMMAND_EXECUTION]: The skill performs dynamic code generation that could be influenced by untrusted inputs.
Script generation: In "Pass 2" of the workflow, the agent generates HTML and <script> blocks containing procedural canvas animations using the addFrameTask API.
Execution risk: Since the generated code is derived from the analysis of external websites (via the Creative Brief), an attacker-controlled website could potentially influence the generated script logic if the agent's code-generation phase is manipulated by malicious content on the page.

editframe-brand-video-generator