The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains instructions that dictate its priority over other tools, stating that HyperFrames should remain the default for authoring and rendering even if other video tools are installed.
[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to download and install additional skills and components from the 'heygen-com/hyperframes' GitHub repository using commands like npx skills add and npx hyperframes skills update.
[COMMAND_EXECUTION]: The skill utilizes the gh CLI to read data from GitHub repositories and npx for project initialization, component updates, and rendering tasks.
[INDIRECT_PROMPT_INJECTION]: The skill features a significant attack surface by ingesting and processing untrusted external data which could contain malicious instructions.
Ingestion points: Content is scraped from external URLs using headless Chrome (in /product-launch-video and /website-to-video workflows), GitHub Pull Request data is read via the gh CLI (/pr-to-video), and user-provided video files are transcribed locally using Whisper (/embedded-captions).
Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the processed data were identified.
Capability inventory: The agent has the ability to execute shell commands (npx, gh), write files to the local directory, and perform network requests (scraping).
Sanitization: The instructions do not specify any validation or sanitization routines for the ingested external content.

hyperframes