The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to fetch and execute the hyperframes package from the npm registry. It also provides commands to manage a pinned version of Chromium (npx hyperframes browser ensure) to ensure reproducible rendering outputs.
[DATA_EXFILTRATION]: The cloud render and publish commands are designed to zip project files—including HTML, scripts, and media assets—and upload them to external infrastructure for cloud processing and hosting. This behavior is the primary intended function of the tool's cloud-based features.
[CREDENTIALS_UNSAFE]: The skill interacts with sensitive authentication tokens and environment variables required for service access. It manages credentials for HeyGen's API and OAuth flows, as well as AWS and GCP credentials for self-managed cloud infrastructure (e.g., ~/.hyperframes/credentials, ~/.aws/credentials).
[INDIRECT_PROMPT_INJECTION]: The skill contains commands that ingest data from untrusted external sources, specifically capture <url> which retrieves content from websites and transcribe which processes audio/video files. These functions represent an attack surface where maliciously crafted external content could potentially influence the agent's behavior during the video creation process.

hyperframes-cli