sources_v4.json
32.4 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
{
"task": "T-mtbn18a4",
"agent": "innovator-hunt (Hunter)",
"round": "retrain v4 HUNT (time-windowed storyboard)",
"crawl_date": "2026-08-27",
"fetched_at_utc": "2026-08-27T14:43:00Z - 2026-08-27T14:56:00Z",
"time_precision_note": "crawl_time 为分钟级精度(同一批次共用一个时刻),非逐请求秒级记录",
"method": "curl --retry 3 --retry-all-errors --max-time 60 直读 raw.githubusercontent.com 全文;24 次请求全部 HTTP 200、零重试零超时;sha256 本地计算;v3 已抓 9 文件全部复抓比对(字节数+sha256 逐一致,见 s01-s09 的 change_vs_v3 字段);行号为文件内 1-based 行号",
"notes": [
"官方 skills/README.md:3 逐字:目录含 1 prompt writing skill + 8 style-specific video generation skills;均为 MiniMax-H3 官方仓库随附内容(A 级官方逐字),但与 h3-prompt-writing 五模式规范属不同层级文本,两种习语的差异见 knowhow_v4.md §2.3",
"Q6 结论:官方上游零变更;本轮全部增量来自 v3 已在树里清点但未读正文的 skills/ 目录 15 个文件"
],
"sources": [
{ "id": "s01", "title": "官方技能总说明 SKILL.md(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/.claude/skills/h3-prompt-writing/SKILL.md", "bytes": 2177, "sha256_16": "a48f9db1231cebe6", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical (bytes + sha256)", "quotes": [ { "line": 33, "text": "- Write rewrite sections in English; preserve dialogue, lyrics, and visible scene text in their original language." } ], "used_for": ["Q6 零变更", "习语甲语言规则(与习语乙冲突点,knowhow_v4 §2.3)"] },
{ "id": "s02", "title": "官方基础模式规范 base-en.txt(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/.claude/skills/h3-prompt-writing/references/base-en.txt", "bytes": 15773, "sha256_16": "2cfebc096a6e0837", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical (bytes + sha256)", "quotes": [
{ "line": 86, "text": "Do not add a timestamp to the first shot. Use sequential shot numbers for later shots, and begin each one with a strictly increasing cut time that falls within the video duration:" },
{ "line": 94, "text": "### 4.3 Camera Motion: Motion Type + Amplitude + Speed" },
{ "line": 96, "text": "A complete camera-motion expression has three dimensions: the motion type defines how the camera moves, amplitude defines the range of compositional change, and speed defines the pacing of that change. Add amplitude and speed only when they are meaningful; medium amplitude and normal speed are usually omitted." },
{ "line": 144, "text": "### 4.5 On-Screen Text" },
{ "line": 146, "text": "Place any banner, sign, label, subtitle, or neon text that is actually visible on screen in English double quotation marks. Preserve the original text and punctuation verbatim, without translation." },
{ "line": 148, "text": "A red neon sign reading \"营业中\" glows above the doorway." }
], "used_for": ["Q2 反证:base-en 无 [Xs-Ys] 窗式写法", "Q3 运镜三维度", "Q3 UI 可见文字条款(v3 未引过 L146)", "Q5 无重叠的相邻条款 L86"] },
{ "id": "s03", "title": "官方 Ref2VA 规范 ref-en.txt(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/.claude/skills/h3-prompt-writing/references/ref-en.txt", "bytes": 23553, "sha256_16": "1e574f356716ad55", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical (bytes + sha256)", "quotes": [
{ "line": 70, "text": "When an image acts as a storyboard or shot-planning reference, state which shots it maps to and what planning information it provides:" },
{ "line": 73, "text": "<Picture 3> is a storyboard reference for [Shot 1] and [Shot 2], defining their viewpoint, subject placement, and shot order." }
], "used_for": ["Q4 分镜参考图映射到镜头而非秒数", "Q4 反证:ref-en 全文 0 处 aligns with(grep 核验)"] },
{ "id": "s04", "title": "openai.yaml(复抓,仅 UI 元数据)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/.claude/skills/h3-prompt-writing/agents/openai.yaml", "bytes": 240, "sha256_16": "7770c9d784b7251a", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical", "quotes": [], "used_for": ["Q6 零变更"] },
{ "id": "s05", "title": "官方仓库 README.md(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/README.md", "bytes": 40135, "sha256_16": "7964f3044d3c7efa", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical (bytes + sha256)", "quotes": [
{ "line": 67, "text": "MiniMax H3 is a general-purpose, omni-modal generative system. It supports unified understanding of multimodal contexts composed of text, images, video, and audio, and can generate video with native stereo audio at resolutions up to 2K and durations of up to 15 seconds." },
{ "line": 73, "text": "| Output duration | 4–15 seconds |" },
{ "line": 326, "text": "- Duration: 10 seconds" },
{ "line": 344, "text": "\"duration\": 10," },
{ "line": 365, "text": "- Duration: 8 seconds" },
{ "line": 404, "text": "- Duration: 5 seconds" }
], "used_for": ["Q1 时长范围官方逐字出处(4–15 seconds)", "Q1 时长由请求体参数指定而非提示词正文"] },
{ "id": "s06", "title": "README.zh-CN.md(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/README.zh-CN.md", "bytes": 36894, "sha256_16": "5c109694dae87cf0", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical", "quotes": [], "used_for": ["Q6 零变更"] },
{ "id": "s07", "title": "Ref2VA processor chat_template.json(复抓,Picture N 自动编号机制性证据)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/Ref2VA/processor/chat_template.json", "bytes": 5499, "sha256_16": "5c72a170d2a4a1a3", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical", "quotes": [], "used_for": ["Q6 零变更", "§9 第 4 条「图1/图2 编号回指与自动编号机制相容」的机制依据(引 knowhow_v3 §2.4)"] },
{ "id": "s08", "title": "官方 Ref2VA 可复现请求脚本(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/scripts/readme/reproducible-768p-ref2va-request.sh", "bytes": 5047, "sha256_16": "73ea0e5cc94440b2", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical", "quotes": [ { "line": 29, "text": "\"duration_seconds\": 5" } ], "used_for": ["Q1 指定方式:target.duration_seconds"] },
{ "id": "s09", "title": "官方 FL2VA 可复现请求脚本(复抓)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/scripts/readme/reproducible-768p-fl2va-request.sh", "bytes": 5718, "sha256_16": "82d1c06897da37e5", "crawl_time": "2026-08-27T14:43:00Z", "change_vs_v3": "identical", "quotes": [ { "line": 25, "text": "\"duration_seconds\": 8" } ], "used_for": ["Q1 指定方式"] },
{ "id": "s10", "title": "仓库全量路径清单 GitHub API(复抓)", "url": "https://api.github.com/repos/MiniMax-AI/MiniMax-H3/git/trees/main?recursive=1", "bytes": 84966, "sha256_16": null, "crawl_time": "2026-08-27T14:44:00Z", "change_vs_v3": "v3 未落盘 tree 全文,无法逐路径历史比对;本轮已落盘摘要,下轮起可比对", "quotes": [ { "line": null, "note": "摘要", "text": "281 条路径。与提示词写作相关的仍为 .claude/.agents/skills/h3-prompt-writing(3 文件×2 副本)、skills/(1 prompt-writing + 8 风格 skill + skills/README.md + skills-lock.json)、scripts/readme 下 18 个脚本。无新增提示词规范文档。" } ], "used_for": ["Q6 无新增规范文件"] },
{ "id": "s11", "title": "官方 skills/README.md(本轮新增,官方定性:1 prompt-writing + 8 style skills)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/README.md", "bytes": 8408, "sha256_16": "d993939102d1b062", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch (v3 清点过路径但未读正文)", "quotes": [
{ "line": 3, "text": "This directory contains the skills bundled with [MiniMax H3](../README.md): **1 prompt writing skill** and **8 style-specific video generation skills**. Each skill lives in its own folder with an installable `SKILL.md` (plus a `SKILL.cn.md` Chinese version for the style skills) and any reference materials it needs." },
{ "line": 7, "text": "The skills are actively maintained and still evolving. The 8 style skills ship with bilingual `SKILL.md`/`SKILL.cn.md`; `h3-prompt-writing` is currently English-only." },
{ "line": 91, "text": "Create a two-player co-op game menu or opening animation. The skill locks identity cues, generates an approval image from a fixed menu framework with coordinated color, buttons, icons, and typography, then uses the approved result to rebuild character, UI-copy, and event timing instructions for the final video." }
], "used_for": ["习语乙的官方定性(A 级)", "knowhow_v4 §1.2/§2.3"] },
{ "id": "s12", "title": "官方 co-op-game-intro-generator/SKILL.md(本轮新增)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/co-op-game-intro-generator/SKILL.md", "bytes": 8216, "sha256_16": "8b49cf84a70fbbc9", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 4, "text": "compatibility: Requires the MiniMax Hub agent (canvas workspace and MiniMax H3 generation); not portable to generic agent harnesses." },
{ "line": 16, "text": "- Use `references/h3-video-prompt-template.md` when refilling the final Minimax H3 video prompt in STEP 6. Fill the template from the approved confirmation image, player/game data, final UI copy, event timing, motion directions, and negative constraints." },
{ "line": 51, "text": "After approval, load `references/h3-video-prompt-template.md` and refill the final video prompt with confirmed style, character refs, player names, game title, UI text, event timing, motion directions, and negative constraints. Generate the final video with Minimax H3." }
], "used_for": ["该模板是直接喂 Minimax H3 的最终 prompt 的官方行为证据(knowhow_v4 §4.5)"] },
{ "id": "s13", "title": "官方 co-op h3-video-prompt-template.md(★本轮核心发现:master v4 案例的官方同型母本,六个时间窗)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/co-op-game-intro-generator/references/h3-video-prompt-template.md", "bytes": 10331, "sha256_16": "7fac20c236de585d", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 18, "text": "## Reference roles" },
{ "line": 19, "text": "- {ui_ref}: confirmed first image. Use it to lock UI layout, menu hierarchy, color system, typography scale, button structure, character-game integration, and overall composition logic." },
{ "line": 20, "text": "- {player1_ref}: PLAYER 1 identity anchor. Lock exact face, hairstyle, glasses if present, facial proportions, body identity, and nickname mapping to {player1_name}." },
{ "line": 22, "text": "- {height_ref}: body comparison anchor. Lock the visible contrast between the two players and prevent identical body proportions." },
{ "line": 39, "text": "视频中的菜单、装备面板、玩家卡、按钮、HUD、加载条、图标和文字都必须沿用同一色彩系统,不得随机新增无关颜色。" },
{ "line": 48, "text": "## Fixed timeline framework" },
{ "line": 50, "text": "### [0秒–2秒] — 双人主菜单" },
{ "line": 51, "text": "景别/机位:高角度俯拍大全景,延续 {ui_ref} 的构图逻辑。镜头从上方轻微下压并缓慢推进。" },
{ "line": 53, "text": "画面内容:PLAYER 1({player1_name})和 PLAYER 2({player2_name})并排坐在画面中央,PLAYER 1 在左,PLAYER 2 在右,抬头看向摄像机。两人只有自然呼吸、眨眼和轻微身体动作。" },
{ "line": 55, "text": "UI结构:左上角玩家资料卡准确显示:" },
{ "line": 64, "text": "右侧纵向菜单准确显示:" },
{ "line": 65, "text": "“START NEW GAME”" },
{ "line": 66, "text": "“CONTINUE”" },
{ "line": 68, "text": "“CONTINUE”是视觉中心和主高亮按钮。" },
{ "line": 70, "text": "动态风格填充:主菜单背景、地面纹理、按钮形态、图标、字体、边框、辉光和贴纸质感全部根据 {visual_style} 优化,但保留 {ui_ref} 的布局和层级。" },
{ "line": 74, "text": "### [2秒–4秒] — PLAYER 1 右臂配置" },
{ "line": 75, "text": "景别/机位:中景,镜头从主菜单平滑推近 PLAYER 1 的右臂,PLAYER 2 仍在背景中可见,不消失、不变形。" },
{ "line": 77, "text": "UI结构:右侧主菜单收缩滑出;带功能强调色识别线的 UI 面板从左侧滑入,准确显示:“PLAYER 1”“RIGHT ARM EQUIPMENT”。装备列表中先高亮“PHANTOM GRIP”,随后选区移动至“CHRONOS CLAW”。" },
{ "line": 79, "text": "动作:PLAYER 1 的右袖口自动打开,轻量机械结构从前臂下方展开。手指分开,修长爪状指节滑入并逐一锁定,内部短暂露出精细线路、微型活塞和金属连接件。配置完成后,功能强调色 LED 依次亮起。" },
{ "line": 85, "text": "### [4秒–7秒] — PLAYER 2 重型手臂配置" },
{ "line": 88, "text": "UI结构:新的暖色/琥珀红识别 UI 滑入,准确显示:“PLAYER 2”“ARMAMENT CUSTOMIZATION”。面板以网格形式展示:" },
{ "line": 101, "text": "### [7秒–8.5秒] — 双人确认配置" },
{ "line": 104, "text": "UI结构:两组装备面板向画面中央汇合,形成共享按钮,准确显示:“CONFIRM CONFIG”。按钮边框、辉光、图标和贴纸质感根据 {visual_style} 优化,但层级清晰、文字可读。" },
{ "line": 106, "text": "动作:光标点击按钮,功能强调色能量脉冲流过 PLAYER 1 的机械爪,暖色/琥珀红能量脉冲流过 PLAYER 2 的机械拳。所有 UI 面板快速向内收缩并消失。" },
{ "line": 110, "text": "### [8.5秒–10秒] — 双人世界加载" },
{ "line": 113, "text": "UI结构:加载条准确显示:“LOADING”。进度从 0% 快速填充至 100%。左半段使用 PLAYER 1 的功能强调色,右半段使用 PLAYER 2 的暖色/力量色。HUD 和加载条的形态、边框、纹理、字体根据 {visual_style} 优化,但必须清晰可读。" },
{ "line": 115, "text": "环境转化:黄色平面环境连续转化为游戏世界。警戒线变成真实街道道路标记和施工围挡;墨迹/纹理化为潮湿路面反光或符合 {visual_style} 的地面光影;平面涂鸦扩展成建筑墙面的喷绘;黑色背景区域形成城市阴影和深巷入口。" },
{ "line": 117, "text": "关键约束:转化必须连续自然,不使用硬切,不用烟雾遮挡,不改变两位角色身份。" },
{ "line": 121, "text": "### [10秒–15秒] — 双人进入游戏世界" },
{ "line": 122, "text": "景别/机位:大全景转第三人称跟拍。加载 100% 的瞬间,两名角色同时起身。摄影机平滑下降并绕到两人身后,变成稳定第三人称双人合作视角。" },
{ "line": 128, "text": "HUD结构:HUD 淡入。右上角出现小地图;左下角出现两组独立状态栏:“{player1_name}”“{player2_name}”。{player1_name} 状态栏使用 PLAYER 1 功能强调色,{player2_name} 状态栏使用 PLAYER 2 暖色/力量色。前方街道中央出现共享任务标记。" },
{ "line": 133, "text": "No third player, no female character unless explicitly requested by user, no duplicated character, no character swapping, no username swapping, no merged bodies, no identical body proportions, no changing faces, no changing hairstyles, no changing identity, no missing PLAYER 1, no missing PLAYER 2, no split screen, no hard cuts, no random camera shake, no floating body parts, no gruesome dismemberment, no weapon, no excessive neon unless selected by user style, no purple background unless selected by user style, no unreadable UI, no random letters, no misspelled usernames, no extra menu options, no official game logo, no copied branded interface, no watermark." }
], "used_for": ["Q2 时间窗分段官方范例(A 级)", "Q3 每窗三要素官方字段(景别机位/动作/UI结构/HUD结构/环境转化)", "Q4 非锚点参考图角色绑定、无秒级对齐", "Q5 六窗构造性首尾相接", "master v4 案例同型判定(knowhow_v4 §2.2 逐项对照)"] },
{ "id": "s14", "title": "官方 co-op h3-confirmation-image-template.md(仅哈希核对,未逐行引用)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/co-op-game-intro-generator/references/h3-confirmation-image-template.md", "bytes": 14001, "sha256_16": "344325132841a673", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [], "used_for": ["确认图工作流留作下轮(knowhow_v4 §12.4)"] },
{ "id": "s15", "title": "官方 3d-animation-short-generator/SKILL.md(本轮新增)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/3d-animation-short-generator/SKILL.md", "bytes": 14961, "sha256_16": "92f1130ccf456b3e", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 12, "text": "Core rule: story first, ask screen size and total duration with choice cards immediately after user intake, ordered canvas artifacts, every required confirmation via choice card, fixed order after character/scene cards: six-column standardized shot table with per-second directives + audio cues + spatial anchor chain → shot-table self-check gate → one single text storyboards document with one section per shot … → video-model choice card (H3 default, Seedance 2.0 fallback) + resolution choice card → single-shot clips rendered by the chosen video model …" },
{ "line": 34, "text": "- Desired output: blueprint only, assets only, standardized shot table with per-second directives, single text storyboards document (default) + extracted single-shot text storyboard nodes for heavy-iteration shots + multi-panel pencil storyboards (opt-in), single-shot video clips (with video model chosen in Step 7), assembled main video, or final BGM-composited film" }
], "used_for": ["Q2 逐秒 panel 制的工作流定位(分镜文档是生产中间产物,供选定的视频模型读取)"] },
{ "id": "s16", "title": "官方 3d-animation storyboard-guidelines.md(本轮新增:panel 结构 + no time gaps 条款)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/3d-animation-short-generator/references/storyboard-guidelines.md", "bytes": 14959, "sha256_16": "a1a5aaed9681dd2f", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 35, "text": "- `Timecode` — e.g. `0–1s`." },
{ "line": 37, "text": "- `Pose + Expression` — concrete body posture, silhouette, key prop grip, eye-line, facial expression path; for elastic beats, explicitly call out squash / stretch / anticipation / overshoot. This is the largest section per panel and is what the video model reads as the visual beat." },
{ "line": 38, "text": "- `Camera` — shot size, camera movement (push / pull / pan / tilt / handheld-shake / locked / orbit), Dutch angle note when applicable." },
{ "line": 42, "text": " - 3-second shot → 3 panels (1 per second)." },
{ "line": 46, "text": " - 7+ second shot → one panel per second; for sub-second critical beats, add an extra mini-panel such as `2.0–2.5s` only when the beat is the hook of the shot." },
{ "line": 47, "text": " - Panels must cover the full shot duration from first frame to last frame with no time gaps." }
], "used_for": ["Q2/Q3 逐秒 panel 三要素", "Q5 无缝隙条款(A 级,习语乙)"] },
{ "id": "s17", "title": "官方 3d-animation shot-table-spec.md(本轮新增:Per-Second Directives 五要素 + no time gap + 15s 上限)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/3d-animation-short-generator/references/shot-table-spec.md", "bytes": 7803, "sha256_16": "d4770ff3f7c6d1d8", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 23, "text": "- **Shot Description**: shot size, camera movement, Dutch-angle design, performance style, SFX, negative prompt, video-model generation notes (the model chosen in Step 7 — H3 or Seedance 2.0 — receives slightly different prompt shapes; for H3 emphasize packaging keywords and text/UI/motion-graphics clarity, for Seedance 2.0 emphasize cinematic camera and elastic performance), and a required `Per-Second Directives` subsection. The subsection must break the shot into second-by-second instructions such as `0–1s`, `1–2s`, `2–3s`; for sub-second critical beats, use `2.0–2.5s` style markers. Each per-second directive MUST cover all five required elements:" },
{ "line": 24, "text": " 1. Action / pose / expression (squash-and-stretch, anticipation, overshoot, follow-through where applicable)" },
{ "line": 25, "text": " 2. Camera movement (push / pull / pan / tilt / handheld-shake / locked / orbit)" },
{ "line": 26, "text": " 3. Spatial position (where the character is, what they hold, what landmark is in frame)" },
{ "line": 27, "text": " 4. Audio cue (narration / dialogue / SFX / breath / silence — or `silent` if intentional)" },
{ "line": 28, "text": " 5. Handoff to the next second or next shot (what state this second locks in for the next one)" },
{ "line": 38, "text": "- Each shot row must include per-second directives that cover the entire shot duration from first frame to last frame, including action, pose, expression, camera movement, spatial position, sound cue, and continuity handoff." },
{ "line": 62, "text": "2. **Single-shot duration**: no shot exceeds 15 seconds. If a beat needs more, split it." },
{ "line": 65, "text": "5. **Per-second directive coverage**: every second from `0s` to the shot duration is covered by a `Per-Second Directives` entry, and each entry contains all five required elements (action/pose/expression, camera, spatial, audio cue, handoff). Sub-second beats like `2.0–2.5s` are allowed but must not leave any time gap." },
{ "line": 66, "text": "6. **Cross-shot continuity**: reading the `Continuity Handoff` column row by row produces a continuous chain — no shot starts from a state that contradicts the previous shot’s ending. Any shot that flips eyeline, character position, prop state, or lighting must mark the flip explicitly (e.g. `HARD CUT — time skip: 2h`)." }
], "used_for": ["Q3 每秒五要素(含 camera + spatial + handoff)", "Q5 无缝隙 A 级条款 + 单镜头 15s 上限", "跨镜头连续性(与 SKILL §5.3 内部增强项相邻的官方条款)"] },
{ "id": "s18", "title": "官方 3d-animation qc-checklist.md(本轮新增,仅关键句)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/3d-animation-short-generator/references/qc-checklist.md", "bytes": 7308, "sha256_16": "ffd9b192f09aff8c", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 37, "text": "- No storyboard artifacts in final video: no panel borders, sketch lines, arrows, labels, handwritten notes, timing marks, pose ghosts, storyboard text, AND no double-binding labels (`[char:…]`, `[scene:…]`, `[shot:…]`, `[dur:…]`, `[hook:…]`)" }
], "used_for": ["分镜标记不进最终成片(官方 QC 条款;对 v4 的启示:分镜/时间窗元数据与交付正文的边界)"] },
{ "id": "s19", "title": "官方 handdrawn-live-video-generator/SKILL.md(本轮新增:固定 0-3/3-6/6-10/10-13/13-15 秒段落序 + 中文优先语言规则)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/handdrawn-live-video-generator/SKILL.md", "bytes": 7911, "sha256_16": "5c62175b552f51ed", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 13, "text": "This Skill **organizes the prompt in the user input language and recommends MiniMax H3 as the confirmed generation step**. Do not generate video until the user explicitly confirms H3 generation. Do not route to planner or executor for prompt-writing. The final prompt language must follow the dominant language of the user input: Chinese input produces Chinese, English input produces English, Japanese input produces Japanese; for mixed input, use the dominant language; when unclear, use the current conversation language. Only user-required proper nouns, model names, or literal parameters may remain unchanged." },
{ "line": 57, "text": "Do not add titles, auxiliary headings such as 舞台设置 or 色味, or explanatory notes. Keep this paragraph order:" },
{ "line": 61, "text": "3. 0-3 秒;" },
{ "line": 62, "text": "4. 3-6 秒;" },
{ "line": 63, "text": "5. 6-10 秒;" },
{ "line": 64, "text": "6. 10-13 秒;" },
{ "line": 65, "text": "7. 13-15 秒;" }
], "used_for": ["Q2 时间窗第三种官方形态(段落序命名制)", "习语乙语言规则与习语甲冲突的直接证据(knowhow_v4 §2.3/§8 判据分叉)"] },
{ "id": "s20", "title": "官方 brand-promo-video-generator/SKILL.md(本轮新增)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/brand-promo-video-generator/SKILL.md", "bytes": 13561, "sha256_16": "a50d270e412be4ed", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 104, "text": "For a 15-second film, target 5-8 major beats. For a 30-second film, target 8-12 major beats. Each beat should define:" }
], "used_for": ["15s 内容的官方切分数参考(5-8 beats,节拍制)"] },
{ "id": "s21", "title": "官方 minimalist-product-ad-generator/SKILL.md(本轮新增)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/minimalist-product-ad-generator/SKILL.md", "bytes": 32495, "sha256_16": "1f07c58090141159", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 14, "text": "The core principle is: **confirm assets and brief first, build product facts and a product narrative spine, lock product visuals with independent anchor photos, control video with a precise beat storyboard, then finish the film with native audio or music-based editing.** This Skill no longer defaults to a 4-panel anchor sheet, because video models may reproduce the panel layout. By default, it uses three separate anchor photos plus one precise beat text storyboard as the video control system." },
{ "line": 60, "text": " - The precise beat storyboard determines the actual shot order, action, and text behavior. Do not use anchor photos as direct shot replacements." }
], "used_for": ["官方把「文字分镜」作为视频控制系统的先例;参考图不替代分镜"] },
{ "id": "s22", "title": "官方 music-video-subtitle-generator/SKILL.md(本轮新增:15s 单次生成上限佐证)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/music-video-subtitle-generator/SKILL.md", "bytes": 14423, "sha256_16": "a2c0904632e8db52", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 34, "text": "7. **Natural multi-shot stitching for >15s videos**: when duration exceeds the model single-generation limit, use “multi-shot storyboard breakdown + tail/head frame continuation + beat hard cuts + global master-audio alignment”. Every segment must preserve the same groove, tempo feel, aspect ratio, character logic, visual preset, lighting language, and typography motion rules." },
{ "line": 63, "text": "3. **30-second or longer complete MV (multi-shot stitching)**: because many models have a 15-second single-generation ceiling, automatically use a Multi-Shot stitching workflow. First generate or lock one continuous song / BGM, split 30 seconds into 4–8 dynamic 2–5 second shots, then stitch via beat sync and head/tail frame continuity." }
], "used_for": ["Q1 时长上限佐证(官方 skill 层);>15s 官方走拆分+续接工作流", "单段 2–5 秒切分的官方先例"] },
{ "id": "s23", "title": "官方 paper-collage-explainer-generator/SKILL.md(本轮新增)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/paper-collage-explainer-generator/SKILL.md", "bytes": 20952, "sha256_16": "2223d4b4ec6b2842", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 85, "text": "For a story topic, split the story into three to six concise beats unless the user specifies a count. Similar beats may share the same design language, but each should have a distinct metaphor, object set, color field, and SFX rhythm." }
], "used_for": ["官方按 beat 切分故事的另一先例(3-6 beats)"] },
{ "id": "s24", "title": "官方 papercraft-stop-motion-explainer/SKILL.md(本轮新增,仅存在性+描述级引用)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/skills/papercraft-stop-motion-explainer/SKILL.md", "bytes": 22882, "sha256_16": "bf8b8e42acd91a02", "crawl_time": "2026-08-27T14:50:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 9, "text": "Create a complete papercraft stop-motion explainer package from a science, education, or knowledge topic. The output is a production-ready creative plan: style rules, character and set design, asset planning, prompts, storyboard structures, motion language, sound direction, negative prompts, and review criteria." }
], "used_for": ["风格 skill 输出形态清点(storyboard 是产物之一)"] },
{ "id": "s25", "title": "官方 T2VA 可复现请求脚本(本轮新增)", "url": "https://raw.githubusercontent.com/MiniMax-AI/MiniMax-H3/main/scripts/readme/reproducible-768p-t2va-request.sh", "bytes": 3425, "sha256_16": "78bfe72365f27a53", "crawl_time": "2026-08-27T14:52:00Z", "change_vs_v3": "new fetch", "quotes": [
{ "line": 19, "text": "\"duration_seconds\": 10" }
], "used_for": ["Q1 指定方式(第三例:T2VA 10s)"] }
],
"negative_findings": [
{ "claim": "base-en.txt / ref-en.txt 不存在 [Xs-Ys] 窗式时间写法", "checked_by": "对两文件全文 grep '\\[[0-9]+[秒s]|[0-9]+-[0-9]+s|[0-9]+秒–' 及逐节通读", "result": "零命中;官方时间表达仅有 [Shot N] 起始时间戳(base-en:86)与首尾帧 aligns with 秒锚(base-en:25/31/201/215)。窗式写法存在于官方 skills/ 目录(s13/s16/s17/s19)" },
{ "claim": "ref-en.txt 全文 0 处 aligns with", "checked_by": "grep -niE 'aligns|align' ref-en.txt", "result": "零命中 → 全参考模式官方没有秒级对齐条款;秒级对齐是 base 模式首尾帧锚点专属(Q4 反证)" },
{ "claim": "gap / overlap 两词在 base-en.txt 与 ref-en.txt 零出现", "checked_by": "grep -niE 'gap|overlap|no time gaps|cover the full' base-en.txt ref-en.txt", "result": "零命中 → 「无重叠/无缝隙」在习语甲中无逐字条款,只有 base-en:86 strictly increasing 相邻条款;无缝隙的逐字条款在习语乙(s16:47 / s17:65)" },
{ "claim": "官方 UI 文字无「中文 (ENGLISH)」中英并列先例", "checked_by": "通读 s13 模板全部 UI 字符串行 + base-en:146 条款", "result": "官方模板 UI 字符串全部英文单语(:65-67 等);master 案例的中英并列属 C 级构造性设计" },
{ "claim": "官方上游本轮零变更", "checked_by": "9 个 v3 文件复抓字节数+sha256 比对 + tree API 重抓清点 281 路径", "result": "全部一致;本轮增量来自 v3 未读正文的 skills/ 15 个文件,非上游新增" },
{ "claim": "「转场」无独立官方字段", "checked_by": "通读 s13 每窗字段集合(景别机位/画面内容/UI结构/动作/动态风格填充/声音/环境转化/关键约束/世界风格/角色关系/HUD结构)", "result": "无转场字段;转场信息落在机位行运镜描述与 Negative constraints 的 no hard cuts(:133)" }
],
"blocked_sources": [
{ "url": "https://platform.minimax.io/docs/api-reference/video-generation-v2-create", "reason": "沿用 v3 §7.1 状态:WebFetch 域安全校验被拒 + curl 超时(本轮未重试,问题集不涉及开放平台 schema)", "impact": "开放平台请求 schema 继续未核实" }
]
}