hyperframes-registry

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches registry manifests and component source files from the HyperFrames repository at https://raw.githubusercontent.com/heygen-com/hyperframes/main/registry. Additionally, the on-device search functionality involves a one-time 33 MB model download.
  • [COMMAND_EXECUTION]: Invokes the hyperframes CLI tool via npx for managing compositions, including searching the catalog, adding items, linting code, and rendering previews.
  • [REMOTE_CODE_EXECUTION]: The skill facilitates the installation and execution of JavaScript-based visual effects (using GSAP and Three.js) within the host environment. This is a core feature of the registry system.
  • [DATA_EXFILTRATION]: Includes a feedback command (hyperframes feedback) that explicitly sends search queries and metadata to the maintainers. This is a manual, user-initiated action designed for reporting search gaps.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external code snippets from a remote registry. While the skill includes instructions for validation (hyperframes lint) and quality checks, the integration of third-party code constitutes an inherent surface for indirect prompt injection or supply chain risks if the source repository were to be compromised.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 01:06 PM