← 返回 Skills 市场

Comfy Story Video

Name: Comfy Story Video
Author: shenlin0517

作者 shenlin0517 · GitHub ↗ · v0.1.0 · MIT-0

cross-platform ⚠ suspicious

337

总下载

当前安装

版本数

在 OpenClaw 中安装

/install comfy-story-video

功能描述

Generate illustrated children's story videos with AI images and TTS narration using ComfyUI running locally.

使用说明 (SKILL.md)

Comfy Story Video

Generate children's story videos with AI-generated illustrations and voice narration.

Quick Start

cd /Users/yiyi/.openclaw/workspace/comfy-story-video
python3 scripts/generate_story_video.py --theme "友谊" --scenes 5

Requirements

ComfyUI running on http://127.0.0.1:8188
Python dependencies: pip install websocket-client requests
FFmpeg for audio conversion

Usage

Generate a Story Video

# Default theme (友谊/friendship)
python3 scripts/generate_story_video.py

# Custom theme
python3 scripts/generate_story_video.py --theme "勇气"

# Custom number of scenes
python3 scripts/generate_story_video.py --theme "环保" --scenes 6

Output Files

Generated in output/ directory:

story_{theme}_{timestamp}.json - Story text
scene_01.png ~ scene_NN.png - AI illustrations
scene_01.mp3 ~ scene_NN.mp3 - Voice narration
story_{theme}_{timestamp}.mp4 - Final video

Customization

Change ComfyUI Model

Edit assets/basic_workflow.json, replace ckpt_name in CheckpointLoaderSimple node.

Modify Characters/Settings

Edit scripts/generate_story_video.py:

characters - List of animal characters
settings - Story backgrounds
themes - Story themes

Change Voice

Modify the -v parameter in generate_tts() function:

say -v '?' | grep zh  # List available Chinese voices

Workflow

Generate story text (configurable characters, setting, theme)
Convert each scene to ComfyUI prompt
Send to ComfyUI for image generation
Generate TTS narration for each scene
Combine into video (WIP)

Troubleshooting

ComfyUI not responding: Check if ComfyUI is running at http://127.0.0.1:8188
Model not found: Update ckpt_name in workflow to match available models
Image generation fails: Check ComfyUI console for error messages

安全使用建议

This package appears internally consistent, but review and caution are still advised before running: 1) It expects ComfyUI on 127.0.0.1:8188 — verify you intend to run that local service and that it is a trusted instance. 2) The included workflow references a checkpoint name (counterfeitV30_v30.safetensors); ensure any model files you use are legally obtained and licensed for your purpose. 3) TTS in the docs uses macOS 'say' and FFmpeg for audio/video work — install appropriate tools for your OS and inspect the script's TTS/ffmpeg invocation (search for subprocess calls) before running. 4) Run the script in an isolated environment (virtualenv or container) if you want to limit impact. 5) If you need higher assurance, open and read the full generate_story_video.py file (and any subprocess calls it makes) to confirm there are no unexpected network calls or shell executions to remote hosts. Overall: technically coherent and limited to localhost operations; use normal precautions (validate models and tools, run in sandbox) before execution.

功能分析

Type: OpenClaw Skill Name: comfy-story-video Version: 0.1.0 The skill contains a significant shell injection vulnerability in 'scripts/generate_story_video.py' due to the use of 'os.system()' to execute 'say' and 'ffmpeg' commands with unsanitized input derived from the '--theme' argument. While the code's logic appears aligned with its stated purpose of generating children's stories via ComfyUI (http://127.0.0.1:8188), the lack of input validation allows for arbitrary command execution if a malicious theme string is provided. No evidence of intentional malice, data exfiltration, or persistence was found.

能力评估

✓ Purpose & Capability

Name/description (generate illustrated story videos with ComfyUI + TTS) match the actual files and runtime requirements. The script communicates with ComfyUI on 127.0.0.1:8188, creates images and audio, and writes outputs to an output/ directory. No unrelated binaries or cloud credentials are requested.

✓ Instruction Scope

SKILL.md and scripts limit actions to: (1) generating story text, (2) converting scenes to ComfyUI prompts, (3) posting to local ComfyUI, (4) retrieving images, (5) producing TTS and combining into video. The docs and code reference local paths and local services only. Note: the README/SKILL.md and code reference macOS 'say' and FFmpeg for TTS/audio processing — these are platform-specific dependencies the user should be aware of.

✓ Install Mechanism

No install spec is provided (instruction-only skill aside from included script), so nothing arbitrary will be downloaded or installed by the skill itself. Dependencies are standard Python packages (websocket-client, requests) and external tools (FFmpeg, optional TTS).

✓ Credentials

The skill requests no environment variables, credentials, or config paths. It only requires a locally running ComfyUI instance and local filesystem access to read assets and write outputs, which is proportional to its purpose.

✓ Persistence & Privilege

always is false and the skill does not request elevated or persistent platform privileges. It does not modify other skills’ configs or request long-lived credentials.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install comfy-story-video
安装完成后，直接呼叫该 Skill 的名称或使用 /comfy-story-video 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v0.1.0

Initial release of comfy-story-video. - Generate illustrated children's story videos using ComfyUI and AI. - Supports custom story themes, number of scenes, and narration. - Outputs story text, AI-generated images, voice narration, and a final video. - Easily modify characters, backgrounds, and voice settings. - Requires ComfyUI, basic Python dependencies, and FFmpeg.

元数据

Slug comfy-story-video

版本 0.1.0

许可证 MIT-0

累计安装 1

当前安装数 1

历史版本数 1

常见问题