Minimax DOCX
poco-ai/poco-claw
Professional DOCX document creation, editing, and formatting using OpenXML SDK (.NET).
用 MiniMax H3（海螺 3.0）生成成片级视频的完整方法论与提示词库。当用户提到「海螺」「Hailuo」「MiniMax H3」「H3 生成视频」「AI 视频提示词」「文生视频/图生视频/首尾帧/全能参考」，或需要把参考片逆向拆解成提示词、要把多段 H3 片段拼成 20–60…
MITAuto-check passed
Install Minimax H3 Video
$ npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install LIUFelix2004/God-minmax-H3 minimax-h3-video --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
Claude Code skills documentation · loads skills from .claude/skills/
Install the "minimax-h3-video" agent skill from https://github.com/LIUFelix2004/God-minmax-H3/tree/main into .claude/skills/minimax-h3-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "minimax-h3-video", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install LIUFelix2004/God-minmax-H3 minimax-h3-video --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "minimax-h3-video" agent skill from https://github.com/LIUFelix2004/God-minmax-H3/tree/main into .agents/skills/minimax-h3-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "minimax-h3-video", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install LIUFelix2004/God-minmax-H3 minimax-h3-video --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "minimax-h3-video" agent skill from https://github.com/LIUFelix2004/God-minmax-H3/tree/main into .cursor/skills/minimax-h3-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "minimax-h3-video", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install LIUFelix2004/God-minmax-H3 minimax-h3-video --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "minimax-h3-video" agent skill from https://github.com/LIUFelix2004/God-minmax-H3/tree/main into .gemini/skills/minimax-h3-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "minimax-h3-video", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install LIUFelix2004/God-minmax-H3 minimax-h3-videoInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "minimax-h3-video" agent skill from https://github.com/LIUFelix2004/God-minmax-H3/tree/main into .github/skills/minimax-h3-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "minimax-h3-video", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install LIUFelix2004/God-minmax-H3 minimax-h3-video --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "minimax-h3-video" agent skill from https://github.com/LIUFelix2004/God-minmax-H3/tree/main into .opencode/skills/minimax-h3-video/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "minimax-h3-video", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
minimax-h3-video- GitHub stars
- 297
- Token cost
- ~647 tokens
- SKILL.md length
- 185 words
- Files
- 18 (incl. scripts, references, assets)
- Skills in repo
- 1
- Repo updated
- First seen
- Licence
- MIT
At a glance
用 MiniMax H3（海螺 3.0）生成成片级视频的完整方法论与提示词库。当用户提到「海螺」「Hailuo」「MiniMax H3」「H3 生成视频」「AI 视频提示词」「文生视频/图生视频/首尾帧/全能参考」，或需要把参考片逆向拆解成提示词、要把多段 H3 片段拼成 20–60…
- Works in 3 steps: 尾帧接首帧 — 上段最后一帧导出为图，作下段 role=first_frame → role=reference_image — 角色定妆图 2–3… → role=reference_audio — 有对白时必传上段音频，否则音色会变
- SKILL.md covers 这个技能解决什么, 先读硬约束, 核心工作流 and 提示词库, plus 2 more sections
- Runs Shell scripts from its folder
What it does
Minimax H3 Video is an agent skill from LIUFelix2004/God-minmax-H3. 用 MiniMax H3（海螺 3.0）生成成片级视频的完整方法论与提示词库。当用户提到「海螺」「Hailuo」「MiniMax H3」「H3 生成视频」「AI 视频提示词」「文生视频/图生视频/首尾帧/全能参考」，或需要把参考片逆向拆解成提示词、要把多段 H3 片段拼成 20–60 秒成片、要保持跨段的角色与风格一致性、要写带运镜与原生音频指令的分镜提示词时，使用本技能。也适用于用户只是发来一段视频问「这是怎么生成的」或「帮我做个 29 秒的片子」。
Its SKILL.md is about 650 tokens, which your agent loads only when the skill is triggered. The skill folder holds 22 other files, including scripts, reference files and assets (for example `PUSH.md`, `README.md` and `prompts/01-final-bet.md`).
It works with MiniMax. The licence is MIT.
Example prompts
- “/minimax-h3-video”
Requirements
- A Bash shell
Workflow steps
3 steps, taken from the first numbered list in SKILL.md.
- 尾帧接首帧 — 上段最后一帧导出为图，作下段 role=first_frame
- role=reference_image — 角色定妆图 2–3 张（正面/侧面/特写）
- role=reference_audio — 有对白时必传上段音频，否则音色会变
What it can do on your machine
Read from SKILL.md and the folder at commit b510e37. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 1 file in scripts/ (Shell, from the files we listed), which the agent can run.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Minimax H3 Video loads about 647 tokens when it runs, and up to ~3.6k if it reads all its reference files. Until then it costs about 61 tokens; SKILL.md has 185 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
SKILL.md
The full file from LIUFelix2004/God-minmax-H3 at commit b510e37, republished under its MIT licence (© LIUFelix2004). 185 words, ~647 tokens.
Download SKILL.mdSave it as
.claude/skills/minimax-h3-video/SKILL.md (or your agent's skills folder). This skill also uses 17 other files; get the full folder from GitHub.- name
- minimax-h3-video
- description
- 用 MiniMax H3（海螺 3.0）生成成片级视频的完整方法论与提示词库。当用户提到「海螺」「Hailuo」「MiniMax H3」「H3 生成视频」「AI 视频提示词」「文生视频/图生视频/首尾帧/全能参考」，或需要把参考片逆向拆解成提示词、要把多段 H3 片段拼成 20–60 秒成片、要保持跨段的角色与风格一致性、要写带运镜与原生音频指令的分镜提示词时，使用本技能。也适用于用户只是发来一段视频问「这是怎么生成的」或「帮我做个 29 秒的片子」。
- license
- MIT
MiniMax H3 视频生成
这个技能解决什么
H3 单次生成上限 15 秒。所有「30 秒 / 1 分钟成片」的需求，本质上都是 分段生成 + 一致性接力 + 拼接三件事。本技能给出这三件事的标准做法， 外加 5 个已验证跑通的分镜提示词模板。
先读硬约束
| 项目 | MiniMax-H3 |
|---|---|
| 单次时长 | 4–15 秒，只接受整数 |
| 分辨率 | 768P / 2K，24fps |
| 提示词上限 | 7000 字符 |
| 全能参考 | 图 ≤9；视频 ≤3（单段 2–15s，总 ≤15s）；音频 ≤3；混合总计 ≤12 个文件 |
| 音频 | 32kHz 立体声与画面同一次生成 |
| 首尾帧 | 0/1/2 张，role=first_frame / role=last_frame |
完整 API 调用见 references/h3-api.md。
核心工作流
第 1 步：判断要不要分段
| 目标时长 | 做法 |
|---|---|
| ≤15s | 一次生成。在一条 prompt 里写多个分镜，H3 会自己切——这是 2.x 时代没有的能力，别浪费 |
| 16–30s | 2 段（15+14 / 15+15）或 3 段（10+10+9） |
| >30s | 按叙事段落切，每段独立可看，别指望长距离一致性 |
**实测结论：**本库 5 个模板里，原本按 3 段（10+10+9）设计的分镜， 直接塞进一条 15 秒 prompt 里生成，H3 完整还原了全部镜头序列和标题卡。 所以优先考虑「一条 prompt 内多分镜」，而不是无脑切段。
第 2 步：写 prompt
结构固定为四块，顺序不要变（详见 references/prompt-grammar.md）：
[风格锚点] 画质/画幅/配色/材质/胶片颗粒 —— 一句话定死整体观感
[分镜序列] 镜头1 → 镜头2 → 镜头3，每个镜头后面挂 [运镜] 标记
[音频指令] 环境声 + 动作音效 + 对白/配乐，写清楚在哪一拍落
[负面约束] 不要出现的东西（多余肢体、错误文字、画面抖动……）风格锚点每一段都要原样重写一遍，不要靠模型跨段记忆。
第 3 步：跨段一致性接力
三件事，缺一件就会飘：
- 尾帧接首帧 — 上段最后一帧导出为图，作下段
role=first_framebash ./scripts/extract_last_frame.sh segA.mp4 segA_last.jpg role=reference_image— 角色定妆图 2–3 张（正面/侧面/特写）role=reference_audio— 有对白时必传上段音频，否则音色会变
第 4 步：拼接
./scripts/stitch.sh 29 out.mp4 segA.mp4 segB.mp4 # 拼接 + 交叉淡化 + 锁死 29.000s原理与逐条命令见 references/stitching.md。
提示词库
prompts/ 下 5 个模板，都是从参考片逆向拆解、再用 H3 跑通验证的。
每个文件含：完整 prompt、API 参数、成片观察、已知坑。
| 文件 | 类型 | 状态 |
|---|---|---|
01-final-bet.md | 红黑白平面矢量动画 / 赌局 | 已验证 15s |
02-hero-vlog.md | 第一人称手持超英 vlog / 电影质感 | 已验证 15s |
03-night-comic.md | 美式硬派漫画 motion comic | 已验证 15s |
04-cavalry-opening.md | 史诗骑队开场 / 长镜 | 待验证 |
05-single-take-dialogue.md | 单镜头对白特写 | 待验证 |
常见失败与对策
| 现象 | 原因 | 对策 |
|---|---|---|
| 出来是固定机位死镜头 | 没写运镜 | 每个分镜后挂 [推进]/[横移]/[跟随] 等标记 |
| 段与段人物长得不一样 | 只靠文字描述 | 必须传 reference_image，文字描述留不住脸 |
| 段与段音色跳变 | 没传参考音频 | 有人声就必须传 reference_audio |
| 画面里出现乱码文字 | 模型自造字 | 负面约束写「不出现任何文字」，标题卡单独做后期 |
| 想要的自创角色被画成某个知名 IP | 描述踩中了训练分布 | 加入明确的差异化特征（配色/材质/剪影），并在负面约束里排除 |
| 拼接处「啪」一声 | 音频硬切 | 用 scripts/stitch.sh，带交叉淡化 |
合规
本库只收录自创设定的提示词。不收录、也不建议写指向具体商业 IP 角色的提示词——那类片子拿不到商用授权，做出来只能自娱自乐。 参考片里若有 IP 角色，拆解时保留其画面语言（配色、运镜、剪辑节奏、 构图法），人物换成自己的设定，风格照样成立。
© LIUFelix2004, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 17 other files (scripts, references, assets) in the repository root of LIUFelix2004/God-minmax-H3.
- SKILL.md
- LICENSE
- PUSH.md
- README.md
- assets/previews/01-final-bet.webp
- assets/previews/02-hero-vlog.webp
- assets/previews/03-night-comic.webp
- prompts/01-final-bet.md
- prompts/02-hero-vlog.md
- prompts/03-night-comic.md
- prompts/04-cavalry-opening.md
- prompts/05-single-take-dialogue.md
- references/h3-api.md
- references/prompt-grammar.md
- references/stitching.md
- scripts/extract_audio.sh
- … and 2 more
Open the folder on GitHubat commit b510e37
Compare with similar skills
Minimax H3 Video next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Minimax H3 Video this skillLIUFelix2004/God-minmax-H3 | 297 | — | ~647 | Automated safety check: Pass | MIT | |
| Minimax DOCXpoco-ai/poco-claw | 1.4k | 7 repos | ~3.9k | Automated safety check: Pass | MIT | |
| AI Image Generation and Editingzhayujie/CowAgent | 47k | — | ~1.3k | Automated safety check: Pass | MIT | |
| Minimax XLSXpoco-ai/poco-claw | 1.4k | 6 repos | ~2.1k | Automated safety check: Pass | MIT | |
| Minimax PDFpoco-ai/poco-claw | 1.4k | 6 repos | ~2.1k | Automated safety check: Pass | MIT | |
| Evals Contextzgsm-ai/costrict | 4.5k | 1 repos | ~1.9k | Automated safety check: Pass | Apache-2.0 |
Similar skills
AI Image Generation and Editing
zhayujie/CowAgent
Generates or edits images from text prompts through a Python script that picks an image backend based on which API keys are configured.
Minimax XLSX
poco-ai/poco-claw
Open, create, read, analyze, edit, or validate Excel/spreadsheet files (.xlsx, .xlsm, .csv, .tsv).
Minimax PDF
poco-ai/poco-claw
A skill your agent uses when visual quality and design identity matter for a PDF.
Evals Context
zgsm-ai/costrict
Provides context about the CoStrict evals system structure in this monorepo.
KrillinAI CLI Operator
krillinai/OpenCreator
Routes agents to the right KrillinAI command for subtitles, dubbing, video rendering, covers and speech, and explains how to read its JSON and manifest output.
Related
Works with
Questions about Minimax H3 Video
What does Minimax H3 Video do?
用 MiniMax H3（海螺 3.0）生成成片级视频的完整方法论与提示词库。当用户提到「海螺」「Hailuo」「MiniMax H3」「H3 生成视频」「AI 视频提示词」「文生视频/图生视频/首尾帧/全能参考」，或需要把参考片逆向拆解成提示词、要把多段 H3 片段拼成 20–60…. Minimax H3 Video is an agent skill from LIUFelix2004/God-minmax-H3.
How do I install Minimax H3 Video in Claude Code?
Run `npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a claude-code`. Or copy the skill folder (the LIUFelix2004/God-minmax-H3 repository) into .claude/skills/minimax-h3-video in your project. Claude Code loads it when a task matches its description.
How do I install Minimax H3 Video in Codex?
Run `npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a codex`. Or copy the skill folder (the LIUFelix2004/God-minmax-H3 repository) into .agents/skills/minimax-h3-video in your project. Codex loads it when a task matches its description.
Can I use Minimax H3 Video in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add LIUFelix2004/God-minmax-H3 --skill minimax-h3-video -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/minimax-h3-video, .gemini/skills/minimax-h3-video, .github/skills/minimax-h3-video and .opencode/skills/minimax-h3-video in your project.
What does Minimax H3 Video need to run?
Going by SKILL.md and its folder, Minimax H3 Video needs a shell for the scripts in its folder. Our summary lists: A Bash shell.
Does Minimax H3 Video access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Minimax H3 Video safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
What licence does Minimax H3 Video use?
Minimax H3 Video is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Minimax H3 Video use?
About 647 tokens (SKILL.md is roughly 2.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.9k tokens, read only when the agent opens those files.
What are the alternatives to Minimax H3 Video?
Skills that share tags, products or a category with Minimax H3 Video: Minimax DOCX (poco-ai/poco-claw, 1.4k stars), AI Image Generation and Editing (zhayujie/CowAgent, 47k stars), Minimax XLSX (poco-ai/poco-claw, 1.4k stars) and Minimax PDF (poco-ai/poco-claw, 1.4k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Minimax H3 Video?
LIUFelix2004 (a GitHub user) maintains it in LIUFelix2004/God-minmax-H3, which has 297 GitHub stars. The repository was last updated on September 1, 2026.
Source: LIUFelix2004/God-minmax-H3 on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.