A command-line quick-reference skill for HyperFrames: lists commands by task for creating projects, checking, viewing frames, editing timelines, previewing, rendering locally or in the cloud, publishing, and viewing history, and teaches agents how to read JSON output and handle build and rendering failures.
skills.sh installations 1285014 (+20521 / 24h)
Security Audit: Risk warnings present
- Installs or runs third-party packages (npx / pip, etc.), which constitutes command execution
- Reports usage data (telemetry)
- Audit scope: GitHub heygen-com/hyperframes/skills/hyperframes-cli/SKILL.md (only scans SKILL.md text)
Rationale
- Complements the main hyperframes skill featured yesterday: the main skill explains how to write compositions, while this one focuses solely on commands; the fixed workflow is lint → check --json to fix all errors → snapshot to read PNGs → preview for human review, with rendering only after human confirmation
- Timeline editing uses the same format as Studio: commands like timeline move / trim / split should first use --plan for preview, each write operation leaves a receipt for undo; multiple changes can be written as JSON and executed at once using timeline apply
- Rendering location is chosen as needed: local, --docker for reproducible containers, batch processing by variable, HeyGen-hosted cloud render, or your own AWS Lambda / GCP Cloud Run
- skills.sh installations in the last 24 hours: 20521, the highest among entries featured this period, with a cumulative total of 1285014; 6 other skills from the same repository are also on the list (indicating attention only)
Usage
- npx skills add https://github.com/heygen-com/hyperframes --skill hyperframes-cli, typically installed alongside the main hyperframes skill
- Requires Node.js 22+ and FFmpeg; commands uniformly use npx hyperframes \u003ccommand\u003e, most support --json, and print a Fix line upon failure
- Cloud rendering requires logging in to your HeyGen account first via npx hyperframes auth login, billed by credits
Risks
- The telemetry mentioned in the audit is real: the author requires running feedback to submit ratings after each successful render, and submitting reports for unmet requirements, unless telemetry is disabled or the user declines; the author states that feedback goes to public channels and requires removing absolute paths and machine identifiers
- The npx execution mentioned in the audit aligns with its purpose; the first run downloads the CLI from npm
- capture \u003curl\u003e fetches external websites; cloud render packages and uploads the entire project to HeyGen