agentCLAWHUBUnverified

Madstory(Seedance Storyboard Engine)

电影级影视分镜设计引擎,支持视频生成(Seedance 2.5默认/2.0降级/Runway/Kling/Sora)及图片生成(Seedream 4.x/5.x)。11种创作模式含短剧全链路。触发:影视分镜/分镜设计/电影分镜/广告分镜/电商视频/UGC广告/品牌短片/多镜头叙事/一镜到底/爆款复刻/短剧创作/AI视频生成/Seedance/Seedream/文生图/图生图/图像编辑/参考图生图/短剧剧本/微短剧/竖屏剧/AI短剧/锁脸/小说改短剧/漫剧/角色人设/时间戳控制/超长视频/视频延长/局部编辑/智能编辑/白模/绿幕/BGM分离/多语种/多宫格分镜。不适用于纯静态视觉设计或非分镜用途 Skill: Madstory(Seedance Storyboard Engine) Owner: qomob Summary: 电影级影视分镜设计引擎,支持视频生成(Seedance 2.5默认/2.0降级/Runway/Kling/Sora)及图片生成(Seedream 4.x/5.x)。11种创作模式含短剧全链路。触发:影视分镜/分镜设计/电影分镜/广告分镜/电商视频/UGC广告/品牌短片/多镜头叙事/一镜到底/爆款复刻/短剧创作/AI视频生成/Seedance/Seedream/文生图/图生图/图像编辑/参考图生图/短剧剧本/微短剧/竖屏剧/AI短剧/锁脸/小说改短剧/漫剧/角色人设/时间戳控制/超长视频/视频延长/局部编辑/智能编辑/白模/绿幕/BGM分离/多语种/多宫格分镜。不适用于纯静态视觉设计或非分镜用途 Tags: latest:3.4.4 Version history: v3.4.4 | 2026-08-

OpenClaw

Rank

62

Safety

84

Downloads

1.7k

Updated

Oct 10, 2026

Version

3.4.4

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.7K downloads reported by the source. Last updated 10/10/2026.

Avoid when

  • Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor
Clawhubvendor · observed Oct 10, 2026
Protocol compatibility
OpenClawcompatibility · observed Oct 10, 2026
Adoption signal
1.7K downloadsadoption · observed Oct 10, 2026
Latest release
3.4.4release · observed Aug 7, 2026
Handshake status
UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s1712hx0t1qgg41g18p2bb0d6583ms02:madstory
  1. Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
  2. Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-qomob-madstory/snapshot"

Documentation

CLAWHUB

149,421 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: mad-story
description: "电影级影视分镜设计引擎,支持视频生成(Seedance 2.5默认/2.0降级/Runway/Kling/Sora)及图片生成(Seedream 4.x/5.x)。11种创作模式含短剧全链路。触发:影视分镜/分镜设计/电影分镜/广告分镜/电商视频/UGC广告/品牌短片/多镜头叙事/一镜到底/爆款复刻/短剧创作/AI视频生成/Seedance/Seedream/文生图/图生图/图像编辑/参考图生图/短剧剧本/微短剧/竖屏剧/AI短剧/锁脸/小说改短剧/漫剧/角色人设/时间戳控制/超长视频/视频延长/局部编辑/智能编辑/白模/绿幕/BGM分离/多语种/多宫格分镜。不适用于纯静态视觉设计或非分镜用途的通用AI绘画。"
version: 3.8.0
author: qomob.ai
license: MIT
modes: 11
phases: 8
platforms:
  - Trae IDE
  - OpenClaw
  - Dify
  - Coze
  - Claude Desktop / MCP
references:
  - references/seedance_25_rules.md
  - references/modes_detail.md
  - references/phases_detail.md
  - references/prompt_engineering.md
  - references/seedance_v2_rules.md
  - references/seedream_4x_rules.md
  - references/terminology.md
  - references/pre_flight_checklist.md
  - references/short_drama_consistency.md
  - references/short_drama_genres.md
assets:
  - assets/cheat_sheet.json
  - assets/storyboard_template.html
---

# MadStory — 电影级影视分镜设计引擎 v3.8.0 (分层加载架构)

## 技能定位

MadStory 是面向电影级内容生产的专业分镜设计引擎,同时支持 **Seedream 4.x/5.x 图片生成**(文生图、图像编辑、参考图生图、多图输入/输出)。

- **触发词**: `影视分镜`, `分镜设计`, `电影分镜`, `广告分镜`, `电商视频`, `UGC广告`, `品牌短片`, `多镜头叙事`, `一镜到底`, `爆款复刻`, `短剧创作`, `从一句话出片`, `AI电影`, `AI视频生成`, `Seedance`, `Seedream`, `分镜脚本`, `视频提示词`, `文生图`, `图生图`, `图像编辑`, `参考图生图`, `短剧剧本`, `微短剧`, `竖屏剧`, `AI短剧`, `短剧分镜`, `短剧编剧`, `锁脸`, `小说改短剧`, `漫剧`, `角色人设`, `时间戳控制`, `超长视频`, `视频延长`, `局部编辑`, `智能编辑`, `白模`, `绿幕`, `BGM分离`, `多语种`, `多宫格分镜`
- **目标用户**: 电影导演、广告导演、创意总监、影视制作人、电商运营、品牌营销、内容创作者、短剧制作人、零基础创作者、平面设计师
- **默认时长**: 30 秒 / 镜头(Seedance 2.5 视频),15 秒(Seedance 2.0 降级时),图片模式无时长限制
- **原生支持平台**: Seedance 2.5(默认)/ 2.0(降级)(视频), Seedream 4.x/5.x/5.0 Pro (图片) — 所有提示词规则、参数体系、质量门禁均基于此二平台校准
- **兼容输出平台**: Runway Gen-3 / Kling / Sora — 输出的提示词可作为创作起点在这些平台使用,但参数体系(Motion Strength、Negative Prompt 格式、多镜头语法等)需用户自行适配

## 规则版本锁定

> **所有提示词规则、参数建议、质量门禁基于以下平台版本验证。** 平台升级后部分规则可能失效,使用前请核对版本。

| 平台 | 验证版本 | 状态 |
|------|---------|------|
| Seedance 2.5 (视频) | **2.5(默认基准)** | 完整规则见 `references/seedance_25_rules.md`;原生30s/超长180s/延长60s/50素材/时间戳/编辑/白模/绿幕/多语种/多人 |
| Seedance 2.0 (视频) | 2.0(降级备选) | 完整规则见 `references/seedance_v2_rules.md`;积分成本敏感场景使用,不支持2.5新增能力 |
| Seedream (图片) | 4.0 / 4.5 / 5.0 Lite / **5.0 Pro** | 全部规则基于此版本范围;5.0 Pro 专属能力见 `references/seedream_4x_rules.md` §8.5 |

### Seedance 2.5 vs 2.0 快速对照

| 维度 | 2.5(默认) | 2.0(降级) |
|------|------------|------------|
| 单次时长 | 30s(可延长至60s,超长模式180s) | 15s |
| 参考图 | ≤30 张 | ≤9 张 |
| 参考视频 | ≤10 段,总≤30s | ≤3 段,总≤15s |
| 参考音频 | ≤10 段,总≤30s | ≤3 段,总≤15s |
| 仅传音频 | 支持 | 不支持(必须配图/视频) |
| 时间戳控制 | 原生秒级/帧级 | 无 |
| 局部编辑/白模/绿幕 | 支持 | 不支持 |
| BGM 分离/多语种/多人 | 支持 | 不支持 |
| 积分消耗(30s 720p) | ~780 | ~210(15s) |

## AI 模型真实限制(必读)

> 以下限制基于 Seedance 2.5 / 2.0 / Seedream 4.x-5.x 的实际生成表现。**2.5 在时长/复杂动作/一致性/Negative 响应/多人/多语种上均有显著改善**。完整对照见 `references/seedance_25_rules.md` §8。

| 限制 | 2.0 表现 | 2.5 改善 | 对策 |
|------|---------|---------|------|
| **时长衰减** | 8-12s 后质量下降,15s 末段纹理崩坏 | 30s 保持一致性 | 2.5 可放心用 30s;2.0 降级仍优先 ≤10s,必须 15s 时标注"8s 后质量风险区" |
| **复杂动作不稳定** | 开门/打斗/转身等多步骤动作易变形 | 长尾动作解锁,

README.md

# MadStory: 电影级影视分镜设计引擎 v3.8.0

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
[![Author: qomob.ai](https://img.shields.io/badge/Author-qomob.ai-blue)](https://qomob.ai)
[![Version: 3.8.0](https://img.shields.io/badge/Version-3.8.0-green.svg)](https://clawhub.ai/qomob/mad-story)

**MadStory** 是一款电影级影视分镜设计引擎。它将模糊的创作意图,通过 8 阶段专业推导流程,逐步转化为包含构图、运镜、光影、表演调度、声音设计等全维度细节的专业分镜提示词。

- **原生平台**: Seedance 2.5(默认)/ 2.0(降级)(视频) + Seedream 4.x/5.x/5.0 Pro (图片)
- **兼容输出**: Runway / Kling / Sora(参数需用户自行适配)

## 与其他 AI 分镜工具的区别

| 维度 | 典型 AI 分镜工具 | MadStory v3.8 |
|------|----------------|---------------|
| **专业理论** | 提示词模板堆砌 | 电影导演工作流:表演调度 + 剪辑理论 + 声音分类论 |
| **知识诚实** | 声称支持所有功能 | 区分 `[模型可执行]` vs `[导演参考]`,不假装 AI 能做它做不到的事 |
| **学习能力** | 静态知识库 | DEFECT_LOG 反馈回路——从每次生成结果中持续优化 |
| **平台诚实** | "支持所有平台" | 原生 vs 兼容分层声明 + 规则版本锁定 + **双版本策略(2.5默认/2.0降级)** |
| **模型限制** | 隐藏缺陷 | 公开声明已知限制及对策,**2.5 vs 2.0 双版本对照** |

## v3.8.0 更新内容

- **新增 Seedance 2.5 全面能力内化(默认基准)**:30s 原生生成、180s 超长视频、60s 视频延长、时间戳帧级控制、50 多模态参考(30图+10视频+10音频)、智能编辑、白模预演、绿幕合成、BGM 分离、多语种(10+ 语种)、多人参考、无缝转场、多宫格分镜
- **新增 双版本策略**:2.5 为默认基准,2.0 降级为成本敏感备选。所有规则/参数/质量门禁基于双版本对照校准
- **新增 2 个创作模式(9→11)**:Mode 9 智能编辑(局部消除/替换/空间视角修改/BGM分离/绿幕)、Mode 10 白模预演(Maya/Blender 粗/细颗粒度渲染)
- **新增 4 段式提示词主公式**:素材描述+一句话概述+具体情节+全局补充。原 5 层结构降级为元素完整性检查清单
- **新增 专项公式索引**:7 维人物公式、30s 3 模块公式、超长视频公式、视频延长公式、智能编辑公式、白模公式、时间戳控制公式
- **新增 seedance_25_rules.md**:2.5 完整规则独立文件(参数规格/16项能力/转场表/Negative禁止项/AI限制2.5改善对照)
- **升级 Negative Prompt**:2.5 用"禁止"表述响应更好(2.0 仍用 `no xxx`)
- **升级 现有 9 个模式**:均新增「2.5 能力增强」小节,标注各自可用的 2.5 新能力

## v3.7.0 更新内容

- **新增 Seedream 5.0 Pro 全部专属能力**: 交互式精准编辑(坐标/框选/标注框)、高密度信息呈现、原生多语种文字渲染、影视感自动增强、人像质感(毛孔级)、Visual CoT 空间推理、多轮对话式编辑、PNG 透明背景输出
- **新增 版本能力对照速查表**: 4.0/4.5/5.0 Lite/5.0 Pro 四版本 14 项能力对比矩阵
- **优化 seedream_4x_rules.md**: 从覆盖 4.0-5.0 Lite 扩展为 4.0-5.0 Pro,新增 §8.5(8 个 Pro 专属子章节)+ §9(版本对照表)+ §10.8(Pro 专属示例)

## v3.6.0 更新内容

- **新增 生成结果反馈回路 (DEFECT_LOG)**: 用户返回生成缺陷 → 按 Mode 累积 → 下次生成自动增强 Negative Prompt 和 KNOWN_RISKS。从静态知识库变成动态经验库
- **新增 AI 模型真实限制声明**: 公开声明时长衰减(8-12s)、复杂动作不稳定、跨镜头一致性靠运气等 5 类已知限制及对策
- **新增 规则版本锁定**: 所有规则标注验证平台版本,平台升级后用户可判断哪些规则仍有效
- **新增 AI 可执行性标注**: Phase 3.5/2/6 每项内容标注 `[模型可执行]` 或 `[导演参考]`,区分"AI 能做到"和"仅供创作决策"
- **新增 Phase 3.5 表演与调度**: 走位(Blocking)、视线匹配(Eyeline)、反应镜头策略、潜台词表演(Subtext)
- **新增 剪辑理论集成**: Phase 2 接入 L-Cut/J-Cut/Jump Cut/Match Cut/Smash Cut/Montage + 景别递进 + 180 度线规则
- **新增 Tournament 评分 Rubric**: 3 维加权(独特性 40% + 一致性 30% + 情感 30%)+ 安全方案惩罚 + 防同质化规则
- **新增 Phase 0 时长充分性检查**: 主动评估 15s 是否够用,不够则建议拆分
- **新增 声音来源分类**: Diegetic / Non-Diegetic / Trans-diegetic 三层决策
- **优化 数据架构**: 删除 5 个冗余 JSON 文件,cheat_sheet.json 成为唯一数据源
- **优化 平台声明**: 区分原生支持(Seedance/Seedream)与兼容输出(Runway/Kling/Sora)
- **优化 短剧工作流**: 明确 Stage A-H 为主线、Phase 0-7 为子流程的映射关系
- **优化 提示词工程文件**: 消除 prompt_engineering.md 与 seedance_v2_rules.md 40% 内容重复
- **优化 清理 5 处 no-op**: 将空泛指导替换为可执行规则

## 11 种创作模式

| 模式 | 适用场景 | 输入方式 | 核心约束 |
|------|---------|----------|------------------|
| **电影创意探索** | 概念开发、风格实验 |

_meta.json

{
  "ownerId": "kn72r3ww47r1qyfaf233sf7q1982rh5f",
  "slug": "madstory",
  "version": "3.4.4",
  "publishedAt": 1786107131713
}

references/examples/example_1_creative_film.md

# 案例 1: 电影创意探索 — 《城市孤岛》概念短片
> **Mode**: `creative_film` (Mode 0) | **平台**: Seedance 3.0 | **时长**: 30s | **风格**: 塔可夫斯基式诗意现实主义

## 用户输入 (Raw Input)

```
我想做一个关于"城市里孤独的人"的短视频,要有电影感,不要太俗套的那种。
```

## 分镜推导过程

### 创意方向筛选

引擎并行生成 3 个创意方向,通过评分矩阵筛选:

| 方向 | 导演参照 | 视觉核心 | 情绪弧线 | 评分 |
|------|---------|---------|---------|------|
| A: 雨夜窗景 | 王家卫 | 湿玻璃反射 + 霓虹虚化 | 孤独→温暖(?) | 6.5 |
| B: 地铁长镜头 | 贾樟柯 | 手持跟拍 + 人群流动 | 疏离→融入→疏离 | 8.2 |
| C: 废墟花园 | 塔可夫斯基 | 固定机位 + 光影流转 | 寂静→顿悟 | **8.8** |

**选定方向**: C(独特性 + 情感深度最优)

### Shot 1: 开场 — 废墟中的光 (0-8s)
- **STANDARD_PROMPT**: `A solitary figure stands motionless in an abandoned rooftop garden overgrown with wild grass, shafts of golden hour sunlight pierce through broken concrete pillars, dust particles float in slow motion, the figure's silhouette barely visible against the intense backlight, cinematic wide shot, static camera, anamorphic lens flare, muted earth tones with selective color on a single red flower`
- **CAMERA**: Static wide shot (固定广角)
- **MOTION_STRENGTH**: 2
- **LIGHTING**: Natural golden hour backlighting through architectural gaps, high contrast between shadow and light pools

### Shot 2: 中段 — 凝视 (8-18s)
- **STANDARD_PROMPT**: `Close-up of weathered hands holding a small cracked mirror, reflecting fragments of the city skyline beyond the garden fence, shallow depth of field, the reflection shifts subtly as wind moves the mirror angle, warm skin tones against cool blue city background, gentle handheld micro-movements`
- **CAMERA**: Extreme close-up with subtle handheld drift
- **MOTION_STRENGTH**: 3
- **LIGHTING**: Mixed warm (skin/mirror) and cool (city reflection)

### Shot 3: 结尾 — 回归 (18-30s)
- **STANDARD_PROMPT**: `The figure now sits on the edge of the rooftop, legs dangling over the cityscape below, camera slowly pulls back (dolly out) revealing the vast urban landscape at dusk, thousands of windows beginning to glow like stars, the figure becomes smaller but no longer alone in frame — the city itself becomes a companion, final frame holds on this composition`
- **CAMERA**: Slow dolly out (缓慢后拉)
- **MOTION_STRENGTH**: 4
- **LIGHTING**: Magic hour transition to blue hour, city lights gradually illuminating

## 完整输出 JSON

```json
{
  "STANDARD_PROMPT": "A cinematic three-shot sequence exploring urban solitude through poetic visual metaphor: [Shot 1] Abandoned rooftop garden, golden hour light through broken concrete, solitary silhouette, static wide shot, anamorphic lens. [Shot 2] Weathered hands holding cracked mirror reflecting city skyline, extreme close-up, shallow DOF, handheld drift. [Shot 3] Figure sitting on rooftop edge, slow dolly out revealing vast dusk cityscape, transition from isolation to cosmic connection.",
  "NEGATIVE_PROMPT": "no generic composition, no cliché visual language, no random style mixing, no emotional disconnect, no flat narrative, no derivative imagery, no shaky camera without motivation, no oversaturated colors, no text overlay, no watermark",
  "TIMELINE": "0-8s: Establ

references/examples/example_2_multi_shot.md

# 案例 2: 多镜头叙事 — 咖啡品牌故事
> **Mode**: `multi_shot` (Mode 4) | **平台**: Runway Gen-3 | **时长**: 3 shots × 5s = 15s

## 用户输入

```
为一个精品咖啡品牌做一支15秒品牌广告,要有温度感,重点展示咖啡豆到杯子的过程。
```

## 分镜方案

### Shot 1 (0-5s)
- **画面**: 咖啡农的手采摘成熟的红樱桃咖啡果,晨光穿过树叶洒在手上,清晰可见的露珠和果实色泽
- **景别**: 特写
- **运镜**: 缓推
- **光影**: 暖金色晨光 (Golden Hour)

### Shot 2 (5-10s)
- **画面**: 新鲜烘焙的咖啡豆从烘焙机中倾泻而出,热气腾腾,近距离展示豆子的油亮表面和色泽
- **景别**: 极近特写
- **运镜**: 微距跟踪
- **光影**: 暖黄工业光

### Shot 3 (10-15s)
- **画面**: 手冲咖啡水流穿过咖啡粉,琥珀色液体滴入杯中,背景虚化中隐约可见咖啡农的微笑
- **景别**: 俯拍特写转中景
- **运镜**: 垂直俯拍到缓拉
- **光影**: 侧逆光突出液体通透感

## 完整输出 JSON

```json
{
  "STANDARD_PROMPT": "[Shot 1] Hands of a coffee farmer picking ripe red cherries, golden morning light filtering through leaves, dewdrops and fruit texture clearly visible, macro extreme close-up, slow push-in, cinematic warm tones. [Shot 2] Freshly roasted coffee beans cascading from roaster, steam rising, oily surfaces glistening, extreme close-up, macro tracking shot, warm industrial lighting. [Shot 3] Pour-over coffee stream through grounds, amber liquid dripping into glass carafe, farmer's smile softly visible in bokeh background, top-down macro to mid-length slow pull-out, side-backlight highlighting liquid translucency.",
  "NEGATIVE_PROMPT": "no inconsistent lighting between shots, no drift in color grading, no mismatch in product appearance, no jumpy transitions, no flat blacks",
  "TIMELINE": "0-5s: Farm harvest; 5-10s: Roasting process; 10-15s: Final brewing",
  "CAMERA": "Slow push-in → Macro tracking → Top-down pull-out",
  "MOTION_STRENGTH": 4,
  "DURATION": "5+5+5=15s",
  "MODE": "多镜头叙事",
  "MODE_KEY": "multi_shot",
  "MULTI_MODAL_ADVICE": "统一色温: 暖色调贯穿三镜头; 建议上传参考图确保咖啡豆/杯具的品牌一致性",
  "SOUND_DESIGN": "Acoustic guitar + ambient farm nature sounds → roasting crackle → pour-over drip",
  "SHOT_LIST": [
    {"shot_id": 1, "duration": "5s", "scene": "coffee_farm_harvest", "focus": "hands_and_cherries"},
    {"shot_id": 2, "duration": "5s", "scene": "roasting_facility", "focus": "beans_cascading"},
    {"shot_id": 3, "duration": "5s", "scene": "brew_bar", "focus": "pour_over_drip"}
  ]
}
```

## Quality Checklist

| 检查项 | 状态 |
|--------|------|
| 镜头数量 ≤ 3 | PASS (3 shots) |
| 跨镜头风格一致 | PASS (统一暖色调) |
| Negative Prompt 完整 | PASS |
| 单镜头单运动原则 | PASS |
Github ReposUpdated 15h agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW
Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW
Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.

MCPOPENCLAW
Github ReposUpdated 7mo agoRank 70

CopilotKit

The Frontend for Agents & Generative UI. React + Angular

OPENCLAW

Machine-readable data

The same record, as JSON, for agents and crawlers.

{
  "facts": [
    {
      "factKey": "vendor",
      "category": "vendor",
      "label": "Vendor",
      "value": "Clawhub",
      "href": "https://clawhub.ai/qomob/skills/madstory",
      "sourceUrl": "https://clawhub.ai/qomob/skills/madstory",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-10T05:31:42.752Z",
      "isPublic": true
    },
    {
      "factKey": "protocols",
      "category": "compatibility",
      "label": "Protocol compatibility",
      "value": "OpenClaw",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-qomob-madstory/contract",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-qomob-madstory/contract",
      "sourceType": "contract",
      "confidence": "medium",
      "observedAt": "2026-10-10T05:31:42.752Z",
      "isPublic": true
    },
    {
      "factKey": "traction",
      "category": "adoption",
      "label": "Adoption signal",
      "value": "1.7K downloads",
      "href": "https://clawhub.ai/qomob/madstory",
      "sourceUrl": "https://clawhub.ai/qomob/madstory",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-10T05:31:42.752Z",
      "isPublic": true
    },
    {
      "factKey": "latest_release",
      "category": "release",
      "label": "Latest release",
      "value": "3.4.4",
      "href": "https://clawhub.ai/qomob/madstory",
      "sourceUrl": "https://clawhub.ai/qomob/madstory",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-08-07T12:52:11.713Z",
      "isPublic": true
    },
    {
      "factKey": "handshake_status",
      "category": "security",
      "label": "Handshake status",
      "value": "UNKNOWN",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-qomob-madstory/trust",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-qomob-madstory/trust",
      "sourceType": "trust",
      "confidence": "medium",
      "observedAt": null,
      "isPublic": true
    }
  ],
  "events": [
    {
      "eventType": "release",
      "title": "Release 3.4.4",
      "description": "Version 3.8.0 introduces major updates with broader platform coverage, more creative modes, and enhanced output specs. - Added native support for Seedance 2.5 (30s默认/超长/局部编辑/多语种/绿幕等) alongside 2.0降级和Seedream 5.0 Pro图片生成。 - 创作模式扩展至11种,涵盖时间戳控制、超长视频、局部/智能编辑、白模、绿幕和多宫格分镜等新场景。 - 输出规范增加ENGINE_VERSION、TIMESTAMP_CONTROL和多平台能力标注,详细区分2.5与2.0参数能力。 - 新增/修订参考文档,如 seedance_25_rules.md、storyboard_template.html,以及丰富的使用示例和细分说明。 - 规则及分层加载结构优化,反映新版平台特性和交互流程。 - 移除旧 skill-card.md,加入丰富的“references”与“assets”文档",
      "href": "https://clawhub.ai/qomob/madstory",
      "sourceUrl": "https://clawhub.ai/qomob/madstory",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-08-07T12:52:11.713Z",
      "isPublic": true
    }
  ]
}

Record generated Oct 10, 2026.

Sponsored

Ads related to Madstory(Seedance Storyboard Engine) and adjacent AI workflows.