Skip to content

gordonlu/awesome-minimax-h3-skills / community-skills/reference-motion-transfer

reference-motion-transfer

Verified

Transfer the motion from a user-supplied reference video onto a new target subject with MiniMax H3. Use when the user wants a character, mascot, product, or creature to perform the dance, gesture, action, locomotion, or camera move they saw in another video. Covers Ref2VA motion transfer (primary route), plus I2VA / FL2VA / L2VA fallbacks when no usable reference video exists. Pure-text requests have nothing to transfer from and are redirected to the base `h3-prompt-writing` skill. Not for style-only transfer with no motion, direct editing of the source video, real-person impersonation without consent, or sequences longer than 15 seconds without segmentation.

FileKindSize
SKILL.cn.mdmarkdown12.1 KB
SKILL.mdskill md13.7 KB
assets/poster.webpbinary30.9 KB
assets/preview.gifbinary2.4 MB
assets/preview.mp4binary1.1 MB
assets/reference.mp4binary2.4 MB
assets/reference.pngbinary1.9 MB
meta.yamlconfig1.8 KB
references/prompt-library.mdmarkdown13.0 KB
display-name-zh: 动作参考迁移
version: 0.1.0
tag-en: Creative & Experimental
tag-cn: 创意实验
complete-tags-en:
- Creative & Experimental / Motion Transfer
- Creative & Experimental / Ref2VA
complete-tags-cn:
- 创意实验 / 动作迁移
- 创意实验 / Ref2VA
summary-en: Transfer the motion from a reference video onto a new target subject with MiniMax H3.
summary-cn: 将参考视频中的动作迁移到全新目标主体上,输出动作一致、身份稳定的新视频。
desc-en: |
  For creators who want a character, mascot, product, or creature to perform the dance, gesture,
  action, locomotion, or camera move seen in a reference video. The Skill analyzes the reference
  motion into a timecoded beat sheet, locks the target subject's identity anchors, builds Ref2VA
  prompts with attribute_transfer retention (with I2VA / FL2VA / L2VA fallbacks), generates
  with MiniMax H3, and reviews motion fidelity, identity stability, and body integrity. Best for
  dance and gesture transfer, mascot performance, and camera-move replication; not for style-only
  transfer, direct editing of the source video, real-person impersonation, or pure-text requests
  without any reference material.
desc-cn: |
  面向想让角色、吉祥物、产品或生物表演参考视频中的舞蹈、手势、动作或运镜的创作者。
  Skill 将参考动作拆解为带时间码的节拍表,锁定目标主体身份锚点,构建带 attribute_transfer
  保留标记的 Ref2VA 提示词(并提供 I2VA / FL2VA / L2VA 兜底路径),用 MiniMax H3 生成,
  最后检查动作保真、身份稳定与身体完整性。适用于舞蹈和手势迁移、吉祥物表演、运镜复制;
  不适用于纯风格迁移、直接剪辑原视频、真人 impersonation,或没有任何参考素材的纯文字请求。
author-en: gordonlu
author-cn: gordonlu
source: community
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall