skills
An AI agent for full-stack short video generation.
Install
mkdir -p .claude/skills/skills-opendemon && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13436" && unzip -o skill.zip -d .claude/skills/skills-opendemon && rm skill.zipInstalls to .claude/skills/skills-opendemon
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
噼哩噼哩 Pilipili-AutoVideo 是一个全自动 AI 视频生成代理,可将自然语言主题转化为完整的短视频(含配音、字幕、转场),并自动生成剪映草稿文件。Key capabilities
- →Generate structured storyboard scripts from natural language topics
- →Generate high-quality keyframe images in parallel
- →Generate TTS voiceovers with precise duration measurement
- →Convert images to video clips using intelligent routing engines
- →Assemble video clips, mix audio, and burn subtitles using FFmpeg
- →Generate JianYing draft files for editing
How it works
Pilipili-AutoVideo is an AI video generation agent that converts natural language topics into short videos by automatically generating scripts, keyframes, voiceovers, and video clips, then assembling them with FFmpeg and creating JianYing draft files.
Inputs & outputs
When to use skills
- →Generating short video content
- →Automating script writing
- →Creating AI-voiced clips
- →Drafting video projects
About this skill
Pilipili-AutoVideo Skill
概述
噼哩噼哩 Pilipili-AutoVideo 是一个全自动 AI 视频生成代理,可将自然语言主题转化为完整的短视频(含配音、字幕、转场),并自动生成剪映草稿文件。
能力
- 从自然语言主题自动生成结构化分镜脚本(LLM 双层架构:生成 + Reflection)
- 并行生成高质量关键帧图片(Nano Banana / Gemini Image Generation)
- 并行生成 TTS 配音(MiniMax Speech-02-HD),精确测量时长
- 图生视频(Kling 3.0 / Seedance 1.5),智能路由引擎选择
- FFmpeg 自动拼接视频、混合音频、烧录字幕
- 生成剪映草稿文件(支持 pyJianYingDraft 或 EDL 回退)
- Mem0 记忆系统:越用越聪明,自动学习用户风格偏好
- 人工审核关卡:脚本生成后暂停,等待用户确认后再调用付费 API
使用方法
方式一:CLI(推荐,快速上手)
# 安装依赖
pip install -r requirements.txt
# 初始化配置
python cli/main.py config --init
# 编辑 configs/config.yaml,填入 API Keys
# 生成视频
python cli/main.py run --topic "AI 改变世界" --style "科技感,蓝紫色调"
python cli/main.py run --topic "西藏旅行" --duration 90 --engine seedance
python cli/main.py run --topic "美食探店" --no-review # 跳过人工审核
# 仅生成脚本(不调用付费 API)
python cli/main.py script --topic "AI 改变世界" --output script.json
方式二:Web UI(可视化操作)
# Docker 一键启动
docker-compose up -d
# 访问 Web UI
open http://localhost:3000
方式三:Python API(程序集成)
from core.config import get_config
from modules.llm import generate_script_sync
from modules.image_gen import generate_all_keyframes_sync
from modules.tts import generate_all_voiceovers_sync, update_scene_durations
from modules.video_gen import generate_all_video_clips_sync
from modules.assembler import assemble_video, AssemblyPlan
from modules.jianying_draft import generate_jianying_draft
config = get_config()
# 1. 生成脚本
script = generate_script_sync(topic="AI 改变世界", duration_hint=60, config=config)
# 2. 并行生成关键帧 + TTS
keyframes = generate_all_keyframes_sync(script.scenes, "output/keyframes", config=config)
audio = generate_all_voiceovers_sync(script.scenes, "output/audio", config=config)
# 3. 更新分镜时长(基于 TTS 实际时长)
script.scenes = update_scene_durations(script.scenes, audio)
audio_paths = {sid: path for sid, (path, _) in audio.items()}
# 4. 图生视频
clips = generate_all_video_clips_sync(script.scenes, keyframes, "output/clips", config=config)
# 5. 组装成片
plan = AssemblyPlan(
scenes=script.scenes,
video_clips=clips,
audio_clips=audio_paths,
output_path="output/final.mp4",
temp_dir="output/temp",
)
assemble_video(plan)
# 6. 生成剪映草稿
generate_jianying_draft(script, clips, audio_paths, "output/draft")
配置说明
配置文件位于 configs/config.yaml(从 config.example.yaml 复制)。
支持的 LLM 提供商(选一个配置 API Key 即可):
- DeepSeek(推荐,性价比高)
- Kimi(月之暗面)
- MiniMax
- 智谱 GLM-4
- Gemini
- OpenAI
- Ollama(本地,无需 API Key)
所有 API Keys 也可通过环境变量设置:
export DEEPSEEK_API_KEY="your-key"
export GEMINI_API_KEY="your-key" # 同时用于图像生成
export MINIMAX_API_KEY="your-key" # 同时用于 TTS
export KLING_API_KEY="your-key"
export KLING_API_SECRET="your-secret"
输出文件结构
data/outputs/{project_id}/
├── script.json # 结构化分镜脚本
├── keyframes/ # 关键帧图片(PNG)
├── audio/ # TTS 配音(MP3)
├── clips/ # 视频片段(MP4)
├── output/
│ ├── {title}.mp4 # 最终成品视频
│ └── 剪映草稿/
│ ├── {title}.draft # 剪映草稿(需 pyJianYingDraft)
│ ├── {title}.edl # EDL 格式(Premiere/DaVinci)
│ ├── {title}.srt # 字幕文件
│ └── {title}_project.json
└── temp/ # 临时文件(可删除)
注意事项
- 首次运行前必须配置 API Keys
- 视频生成(Kling/Seedance)耗时较长(每段约 1-3 分钟),请耐心等待
- 支持断点续传:已生成的文件不会重复生成
- 记忆系统数据存储在
data/memory/mem0.db,可随项目迁移
When not to use it
- →When the user needs to manually edit video content from scratch
- →When the user wants to generate videos outside of short video format
- →When the user needs to use video generation engines not supported by the skill
Limitations
- →Requires API Keys for LLM, image generation, and TTS providers
- →Video generation (Kling/Seedance) can be time-consuming
- →Supports breakpoint resume, so generated files are not regenerated
How it compares
This skill automates the entire short video creation process from a natural language topic, including scriptwriting, image and audio generation, video assembly, and draft file creation, which differs from manual video production workflows.
Compared to similar skills
skills side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| skills (this skill) | 0 | 4mo | Review | Intermediate |
| sora | 2 | 6mo | Review | Intermediate |
| promo-video-builder | 0 | 2mo | Review | Advanced |
| video-downloader | 101 | 7mo | Review | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
sora
davila7
Use when the user asks to generate, remix, poll, list, download, or delete Sora videos via OpenAI’s video API using the bundled CLI (`scripts/sora.py`), including requests like “generate AI video,” “Sora,” “video remix,” “download video/thumbnail/spritesheet,” and batch video generation; requires `OPENAI_API_KEY` and Sora API access.
promo-video-builder
luoxiaojun1992
|
video-downloader
ComposioHQ
Downloads videos from YouTube and other platforms for offline viewing, editing, or archival. Handles various formats and quality options.
annas-archive-ebooks
ratacat
Use when needing to look up book content, find a book by title/author, download an ebook, or reference material from a published book. Triggers on book lookups, ebook downloads, "find the book", "get the PDF/EPUB of". Downloads produce PDF/EPUB/MOBI files - use ebook-extractor skill to convert to text.
openai-whisper
openclaw
Local speech-to-text with the Whisper CLI (no API key).
jianying-editor
luoluoluo22
剪映 (JianYing) AI自动化剪辑的高级封装 API (JyWrapper)。提供开箱即用的 Python 接口,支持录屏、素材导入、字幕生成、Web 动效合成及项目导出。