← 返回 Skills 市场

抖音民间诡异故事视频生成器

Name: 抖音民间诡异故事视频生成器
Author: google696

作者 google696 · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

336

总下载

当前安装

版本数

在 OpenClaw 中安装

/install douyin-folklore-video

功能描述

抖音民间诡异故事视频生成器。根据用户提供的文案自动生成配音、图片、视频并合成最终视频，支持上传到抖音。触发词：生成抖音视频、民间故事视频、诡异故事视频、AI视频制作。

使用说明 (SKILL.md)

抖音民间诡异故事视频生成器

根据用户提供的文案，全自动生成抖音风格的民间诡异故事视频。

完整流程

1. 获取故事文案

必须使用用户提供的文案
保存到 story.txt

2. 生成配音

调用 scripts/generate_voice.py：

python scripts/generate_voice.py \x3C故事文案> \x3C输出目录>

输出：voice.wav
返回：配音时长（秒）

3. 生成图片

调用 scripts/generate_images.py：

python scripts/generate_images.py \x3C故事文案> \x3C输出目录>

根据故事生成 5-6 张图片
尺寸：720×1280
风格：dark horror comic style
保存：img1.png ~ imgN.png

4. 生成视频段

调用 scripts/generate_videos.py：

python scripts/generate_videos.py \x3C图片URL列表> \x3C输出目录>

每段 5 秒
按顺序推进（不回退）
保存：v1.mp4 ~ vN.mp4

5. 视频拼接

ffmpeg -f concat -safe 0 -i list.txt -c copy video_raw.mp4

6. 生成同步字幕

调用 scripts/generate_subtitle.py：

python scripts/generate_subtitle.py \x3C配音时长> \x3C故事文案> \x3C输出目录>

根据配音时长精确计算每句字幕时间
输出：subtitle.srt

7. 视频合成

# 加字幕
ffmpeg -i video_raw.mp4 -vf "scale=540:960,subtitles=subtitle.srt" -c:v libx264 -crf 23 -an video_sub.mp4

# 加音频
ffmpeg -i video_sub.mp4 -i voice.wav -c:v copy -c:a aac -map 0:v -map 1:a final_video.mp4

8. 发送给用户审核

用 message 工具发送 final_video.mp4 给用户。

9. 用户确认"发布"后上传抖音

使用 browser 工具：

profile: openclaw
上传文件到 C:\Users\SS\AppData\Local\Temp\openclaw\uploads\
找到 input[type=file] 元素上传
填写标题、描述、话题
添加 AI 声明
点击发布

关键参数

参数	值
配音模型	qwen3-tts-flash
图片模型	wan2.6-t2i
视频模型	wan2.2-kf2v-flash
图片尺寸	720×1280
视频分辨率	540×960
字幕字号	12
视频编码	H264

配置信息

见 references/config.md：

API Key
FTP 服务器信息
浏览器配置

视频结构规则

按顺序推进，不回退：

v1: img1 → img2
v2: img2 → img3
v3: img3 → img4
v4: img4 → img5
...

视频时长规则：

配音时长 + 1~2秒 = 视频时长
每段视频固定 5 秒
根据配音时长计算需要多少段视频

注意事项

字幕必须根据配音时长精确计算，确保同步
视频要比配音长 1-2 秒
上传抖音时必须添加 AI 声明
用户确认"发布"后才能点发布按钮

安全使用建议

Before installing or running this skill, verify the following: (1) Ask the publisher to update the registry metadata to explicitly list required environment variables (DASHSCOPE_API_KEY, FTP_HOST/USER/PASS/BASE_URL, any Douyin account info) so you know what secrets you'll need to provide. (2) Confirm that dashscope.aliyuncs.com is the intended, trusted API provider and that you trust that provider with your text and media. (3) Carefully control the FTP server: images are uploaded to the configured FTP and will be publicly reachable via FTP_BASE_URL — use a server you control or a temporary/test server. (4) Be cautious with the browser profile approach: uploading via a local profile may use an already logged-in Douyin account; avoid using your primary account or supply an isolated browser profile. (5) Run first in an isolated environment (VM/container) and with test accounts/keys. (6) If you require higher assurance, request that the author supply minimal reproduction steps and explicit listing of required secrets, or review and possibly modify the scripts to avoid exposing media to third-party servers. The SKILL.md states 'user confirmation before publish' — ensure the agent truly waits for interactive confirmation and does not autonomously publish.

功能分析

Type: OpenClaw Skill Name: douyin-folklore-video Version: 1.0.0 The skill bundle automates video creation and social media uploading, but exhibits several high-risk patterns. Specifically, 'scripts/generate_videos.py' requires FTP credentials to be stored in environment variables and uses the insecure FTP protocol to host images for an external API. Additionally, 'SKILL.md' contains hardcoded absolute file paths (e.g., 'C:\Users\SS\AppData\Local\Temp\openclaw\uploads\') and instructs the agent to perform automated browser actions on Douyin (TikTok), which is a high-privilege activity. While these actions appear aligned with the stated purpose, the combination of insecure data handling and specific environment assumptions warrants a suspicious classification.

能力评估

⚠ Purpose & Capability

The description matches the included scripts (voice, image, video, subtitle generation and assembling). However the registry metadata declares no required environment variables or credentials, while the code clearly requires DASHSCOPE_API_KEY and FTP credentials (FTP_HOST, FTP_USER, FTP_PASS, FTP_BASE_URL). The SKILL.md also references Douyin account and browser profile configuration. The missing declared env vars/config in metadata is an incoherence.

⚠ Instruction Scope

Runtime instructions are explicit and scoped to the stated task, but they require writing files to a local temp uploads directory and using a browser profile named 'openclaw' to upload to Douyin. The SKILL.md tells the agent to upload images to an external FTP (for the video-generation API to consume). Those actions expose local files and rely on local browser profiles/accounts; the instructions do not ask for or validate that these external endpoints/profiles are trusted.

ℹ Install Mechanism

There is no install spec (instruction-only), which reduces installer risk. However the skill bundles executable Python scripts which will be executed at runtime and make network calls to dashscope.aliyuncs.com and arbitrary FTP hosts — so there is runtime network exposure even without an install step.

⚠ Credentials

The code requires sensitive environment variables (DASHSCOPE_API_KEY and FTP credentials) and references Douyin account/profile values in the config doc, yet the skill metadata lists none. Requiring an API key and FTP credentials is plausible for this functionality, but the omission from metadata and the number of credentials required (including potentially a logged-in browser profile) is disproportionate to what the registry advertises and should be declared.

✓ Persistence & Privilege

always is false and the skill does not request persistent installation or modify other skills. It uses a browser tool/profile for uploading; autonomous invocation is allowed by default, which increases blast radius if combined with other issues, but that alone is not a new mismatch.

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 douyin-folklore-video
安装完成后，直接呼叫该 Skill 的名称或使用 /douyin-folklore-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

v1.0.0

首发版本：自动生成配音、图片、视频并合成，支持上传抖音

元数据

Slug douyin-folklore-video

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题