hyperframes-cli
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill includes a feedback mechanism that sends reproduction packets to a public channel. It provides explicit instructions for users and agents to redact absolute paths, machine identifiers, and credentials before submission.
- [INDIRECT_PROMPT_INJECTION]: The
capturecommand ingests live URLs to scaffold editable projects, creating a vulnerability surface for untrusted web content. The skill defines a 'hard stop' protocol if the capture identifies blocked content or safety violations, preventing downstream actions like rendering from being performed on malicious inputs. - [EXTERNAL_DOWNLOADS]: The skill performs remote downloads for AI models (Whisper, ONNX), framework components from the HyperFrames registry, and additional framework skills from GitHub, all of which are essential to its documented video processing functionality.
- [COMMAND_EXECUTION]: The CLI orchestrates complex local and remote operations, including FFmpeg processing, Chromium browser management, and the deployment of serverless infrastructure using AWS SAM and Terraform.
Audit Metadata