website-to-hyperframes
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill captures and processes content from external websites provided by the user, which is a significant surface for indirect prompt injection. Malicious instructions placed on a target website could attempt to manipulate the AI's script-writing, design decisions, or command parameters.
- Ingestion points: Website data is extracted and read in Step 1 (Capture & Understand).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used to isolate the captured data from the agent's primary instructions.
- Capability inventory: The skill has the capability to write multiple project files (DESIGN.md, SCRIPT.md, STORYBOARD.md, etc.) and execute the
hyperframesCLI tool via shell. - Sanitization: No sanitization or content-filtering steps are defined for the website data before it is utilized in subsequent workflow steps.
- [COMMAND_EXECUTION]: The workflow relies on
npx hyperframescommands to lint and validate the generated video compositions and project artifacts.
Audit Metadata