Docs & Knowledge · heygen-com/hyperframes

hyperframes

Name: hyperframes
Author: heygen-com/hyperframes

Mandatory entry point: read this first for any request to make, create, edit, animate, or render a video, animation, or motion graphic, including a promo, explainer, captioned clip, title card, overlay, slideshow or interactive deck, Remotion port, or any HyperFrames HTML composition. Also use it to inspect, diagnose, validate, preview, publish, or batch-render an existing HyperFrames project. Inputs may be a website URL, GitHub PR, Figma design or URL, text or brief, existing footage, or music. It resumes project state, captures intent when applicable, selects and installs the owning workflow, and routes domain capabilities. HyperFrames is the default output framework unless the user explicitly chooses another framework for the deliverable or asks only to record a browser session.

L4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandExternal WriteNetwork AccessWrite FileDual-pass Audit

Author heygen-comGitHub heygen-com/hyperframes ↗Stars 48297License Apache-2.0（仓库根 LICENSE；GitHub API spdx Apache-2.0）Commit b8328f9573

Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Acts as the mandatory entry point and routing layer across the 20+ skill ecosystem: SKILL.md uses a table-driven decision process based on existing project state (port / specific operation / edit / BRIEF.md / hyperframes.json / fresh creation), stopping at the first matching row instead of re-running intent interviews every time.

skills/hyperframes/SKILL.md

Apply the first matching row; do not evaluate lower state rows:

The state table consists of 6 rows (Remotion port → specific operation → specific edit → BRIEF.md exists → resume from existing project files → Fresh creation run intent layer); 'If a fresh request does not identify the subject or input, ask what the video is about before routing.'

Directs fresh creations to a dedicated intent interview layer (references/intent-interview.md 8 steps: memory → triage → pick route → pitch → must-haves → run-shape → write BRIEF.md → wrap-up), crystallizing interview findings into BRIEF.md—making the brief the sole routing artifact, after which this skill is never re-read.

skills/hyperframes/SKILL.md

For fresh creation the intent layer (`references/intent-interview.md`) runs the full conversation — memory, triage, pitch round, must-haves, run-shape, hand-off — and **ends by writing `BRIEF.md`. The brief is the only routing artifact the workflow reads**

intent-interview.md specifically includes 8 steps: 1 memory before questions (runs media-use prefs/recipe scripts to read user memory), 2 triage, 3 pick route, 4 pitch round (presents 5 concepts for unformed requests), 5 must-haves, 6 two run-shape questions (storyboard? / automation vs companion), 7 brief, 8 wrap-up.

Matches distribution routes by deliverable rather than keywords: 10 workflow rows (remotion port → slideshow → embedded-captions → talking-head-recut → music-to-video → motion-graphics → pr-to-video → product-launch-video → faceless-explainer → general-video) accompanied by 7 disambiguation rules.

skills/hyperframes/SKILL.md

Use the first matching row. Match the requested **deliverable**, not a word or file type mentioned in passing.

Requires reading references/routes/<workflow>.md before routing (a ~1KB micro-file per route defining I/O, trigger contracts, and the route's interview entry point); if a candidate fails the contract, routing continues rather than forcing a fit.

Mandates refreshing or installing target workflows and core domain skills prior to entering a workflow via `npx hyperframes skills update <workflow-name>`; fails hard on error and prohibits reconstructing workflows from memory, enforcing the CLI as the single source of truth.

skills/hyperframes/SKILL.md

Before reading the selected workflow, install or refresh it and the core domain skills: ```bash npx hyperframes skills update <workflow-name> ``` Use the bare name without `/`. If the command fails, surface the error; do not reconstruct the workflow from memory.

Installation mechanics (core-vs-lazy splitting, init refresh, CI opt-out, non-CLI fallback) are documented in references/skill-lifecycle.md—this skill is strictly responsible for selecting the correct workflow and ensuring its presence.

Provides an on-demand domain skill loading table (composition structure → core, motion → animation, keyframes, creative, media-use, audio, cli, registry, figma) with a precise loading matrix for cross-domain requests like 'Creator edit phrases'; domain skills never assume ownership of end-to-end deliverables.

skills/hyperframes/SKILL.md

| zoom in here, punch-in / punch-out, smooth multi-state zoom or reframe, Ken Burns, or camera move | `/general-video` + `/hyperframes-core` + `/hyperframes-keyframes`; animate the inner visual/crop wrapper, not the timed clip. |

The matrix maps creative verbs (cut/zoom/transition/fade/mix/combinations) precisely to 'load A+B+C' combinations to prevent under-loading or over-loading; 'Domain skills never take ownership of the end-to-end deliverable.'

Integrates version discipline: checks for outdated pins upon resuming a project via a read-only probe (`npx hyperframes@latest upgrade --project . --check`), upgrades if outdated, and verifies using `npx hyperframes check`—the CLI version pin is owned by project package.json scripts and never drifts automatically.

skills/hyperframes/SKILL.md

A scaffolded project pins `hyperframes@<version>` in its `package.json` scripts so renders stay reproducible; the pin never advances on its own, and a pinned run of an older CLI prints no warning about it.

The probe command explicitly specifies keeping the explicit `.` (older CLI versions swallow flags following a bare --project as directory values)—such pitfall-level guidance demonstrates this skill was refined through extensive real-world usage.

2Core Capabilities

01Project-state-aware session resumption (resume/edit/single operations without re-running interviews)

02Intent capture: 8-step interview converging 'make me a video' into an actionable BRIEF.md (memory-first, pitch round, must-haves, run-shape)

0310-way workflow routing + disambiguation (beat-synced vs background music, title cards vs motion-graphics, storyboard as a process rather than a workflow, etc.)

04On-demand workflow installation/refresh (`skills update`) with explicit failure propagation

05On-demand domain skill loading matrix (including cross-domain combinations for creator editing phrases)

06CLI version pin detection and upgrade validation (`upgrade --check` + `check`)

3External Dependencies

Kind	Dependency / Endpoint
cli	hyperframes CLI（skills update / upgrade --check / 项目内 npx hyperframes）
cli	media-use 的 prefs.mjs/recipe.mjs（意图层 step 1 读用户记忆/配方；intent-interview.md 指引）
network	（间接）npx 首次拉取 / skills update 与 GitHub 最新版比对

4Recommended Governance & Evidence

Suggested Controls for Safe Operation

For Agent Runtime

Graded/per-use authorization, sandbox recommended

For Administrators

Approval workflow, rollback support, full audit

Evidence & Triggers (File Paths & Excerpts)

scripts/: Execute local scripts or commands: node <MEDIA_DIR>/scripts/prefs.mjs get / recipe.mjs list (reads user preferences/recipes, read-only)
外部 API: Publish posts, create resources, or modify data on external platforms
环境变量 / API Key: Read vendor service API keys: intent layer reads user memory (prefs.mjs/recipe.mjs access project .media/preferences.json, recipes/, and ~/.me)
本地工作目录: File read/write and artifact output: writes BRIEF.md (interview deliverable, the sole routing artifact for workflows)

Audit Notes: This skill is a pure orchestration and routing layer: it only writes BRIEF.md, reads local preference files (prefs/recipe containing creative preferences, no credentials), and runs official hyperframes CLI commands (update/upgrade/check pulled via official npm channels without handling user quotas or secrets—actual rendering, asset operations, and credentials are delegated to downstream workflows/domain skills such as media-use). It reads no credentials and performs no third-party outbound transmissions; network access during skills update/upgrade routes through official distribution channels, categorized on par with local CLI workflows. Borderline network behavior (npx package fetching) is interpreted conservatively under uniform grading as standard official CLI usage without escalating risk levels.

5Second-Pass Verification

[ok] 『入口/路由』定位与 README/生态描述一致 — README.md：'Read `/hyperframes` first — it's the router and capability map'；CLAUDE.md/AGENTS.md 亦把 hyperframes 列为入口（仓库级 grep 复核）。
[ok] 八步访谈与 step 1 读记忆脚本调用 — intent-interview.md 原文 'Run `node <MEDIA_DIR>/scripts/prefs.mjs get --hyperframes <MEMORY_ROOT> --json`' 复核一致；prefs.mjs/recipe.mjs 存在于 media-use/scripts/。
[ok] 是否有隐藏脚本/网络（纯文档声称） — find 实测 skills/hyperframes 仅 SKILL.md + references/（10 文件 + routes/ 9 文件），零可执行文件；URL 扫描零第三方域命中。
[discrepancy] BRIEF.md 指令载体风险是否遗漏 — 首轮未单列『BRIEF.md 作为可信指令』的注入视角；二轮补充进 injection_surface：项目内既有 BRIEF.md 会被直接执行其 workflow/flow，来源不可信时属项目文件信任面。
[ok] 工作流/域技能不越权承诺（end-to-end 归属） — SKILL.md 'Domain skills never take ownership of the end-to-end deliverable. Load only what the active workflow needs.' 与加载矩阵共存，职责边界自洽。

6Verdict & Scenarios

Architects the entry point across state machine, routing, and intent layers—resumes existing projects without redundant interviews and minimizes questions for new projects to optimize interaction costs

Establishes BRIEF.md as the single source of truth, eliminating cross-skill information loss and maintaining an auditable pipeline

Maps cross-domain loading matrices precisely to creative action verbs, avoiding both skill omission and kitchen-sink loading

Builds CLI version pin detection and upgrade verification into resumption flows, making rendering reproducibility the default

Zero-code, zero-credential pure orchestration implementation with minimal attack surface and full directory portability

Best for: Ideal as the entry session for any HyperFrames video creation task in skill-enabled hosts (Claude Code, Cursor, Gemini CLI, Codex, etc.), as well as for practitioners studying multi-skill ecosystem architectures using a routing layer for state resumption, intent capture, on-demand installation, and domain routing.

Not for: Not suited for low-level creative execution details (it only routes and does not implement—delegate rendering, animations, or asset handling to respective domain skills); not suited for non-HyperFrames video workflows (it is not a generic video assistant, strictly serves the HTML-to-video framework, and explicitly notes 'HyperFrames is the default output framework unless the user explicitly chooses another').

Install Agent-Ready

① Verified Mirror 2026-09-09

Option A · Direct DownloadDownload hyperframes.tar.gz

sha256: 90b6328fc5285f0f…

Option B · Copy Guide for Agent

install.json

Agents automatically download and verify mirrors from this JSON guide.

② Upstream GitHub

GitHub Repo ↗

Pinned to commit b8328f9573.