← 返回 Skills 市场
694
总下载
0
收藏
4
当前安装
1
版本数
在 OpenClaw 中安装
/install douyin-download-v2
功能描述
抖音无水印视频下载和文案提取工具
使用说明 (SKILL.md)
douyin-download Skill
抖音无水印视频下载和文案提取工具。
功能
- 🎬 获取无水印视频下载链接
- 📥 下载抖音视频
- 🎙️ 从视频中提取语音文案(需要 API Key)
- ✂️ 语义分段(调用 OpenClaw 内置 LLM)
环境变量
SILI_FLOW_API_KEY- 硅基流动 API 密钥(用于语音转文字)
获取 API Key: https://cloud.siliconflow.cn/
使用方法
获取视频信息
node /root/.openclaw/workspace/skills/douyin-download/douyin.js info "抖音分享链接"
下载视频
node /root/.openclaw/workspace/skills/douyin-download/douyin.js download "抖音链接" -o /tmp/douyin-download
提取文案(自动语义分段)
node /root/.openclaw/workspace/skills/douyin-download/douyin.js extract "抖音链接"
- 自动调用 Silicon Flow ASR 提取文字
- 自动调用 OpenClaw 内置 LLM 进行自然语义分段
跳过语义分段
node /root/.openclaw/workspace/skills/douyin-download/douyin.js extract "抖音链接" --no-segment
安全使用建议
Before installing or supplying credentials, consider the following:
- Missing declarations: the skill requires ffmpeg (declared) but the code also uses 'curl' (spawned) and an optional MINIMAX_API_KEY — 'curl' and MINIMAX_API_KEY are not declared in the skill metadata. Ask the publisher to add MINIMAX_API_KEY to requires.env and declare 'curl' as a required binary or change to a pure-Node HTTP implementation.
- Credential scope: SILI_FLOW_API_KEY will be sent to https://api.siliconflow.cn with your audio file. If you set MINIMAX_API_KEY it will be sent to https://api.minimaxi.com. Only provide API keys you trust and that are scoped/limited; avoid sharing high-privilege or billing-sensitive keys.
- Network & file-exfiltration risk: the skill uploads audio files to external services (ASR and segmentation). While that matches the declared purpose, it means sensitive audio/text could be transmitted. If you must process sensitive content, run the skill in an isolated environment (container or VM) or review/modify the code to remove external calls.
- Behaviour mismatch: SKILL.md claims use of an internal LLM for segmentation but the code calls MiniMax. Confirm with the author which model/service will actually be used.
- Operational recommendation: run the script in a sandboxed environment first, inspect/verify the code yourself (or ask for a signed/published source), and monitor outbound network requests during a test. If you don't want external calls beyond Silicon Flow, remove or disable the MiniMax call.
- Legal/ethical note: downloading videos from platforms may violate terms of service or copyright — ensure you have rights to download and process the content.
If you want, I can produce a short patch suggestion to (a) remove undeclared 'curl' usage by using Node HTTPS for API calls, (b) add explicit checks and warnings when MINIMAX_API_KEY is missing or present, and (c) surface required binaries/env variables clearly in SKILL.md.
功能分析
Type: OpenClaw Skill
Name: douyin-download-v2
Version: 2.0.0
The skill `douyin-download-v2` is designed to download Douyin videos, extract audio, transcribe speech using SiliconFlow API, and optionally segment text using MiniMax API. The `SKILL.md` provides clear instructions without any prompt injection attempts. The `douyin.js` script uses `child_process.spawn` with an array of arguments for `ffmpeg` and `curl` commands, which significantly mitigates shell injection risks. It accesses `SILI_FLOW_API_KEY` and `MINIMAX_API_KEY` from environment variables for legitimate API calls to `api.siliconflow.cn` and `api.minimaxi.com`, which aligns with its stated purpose. File operations are confined to `/tmp` or user-specified directories, and there is no evidence of data exfiltration to unauthorized endpoints, persistence mechanisms, or other malicious behaviors.
能力评估
Purpose & Capability
The name/description (Douyin no-watermark download + text extraction) matches the code: it fetches Douyin pages, extracts a no-watermark video URL, downloads video, extracts audio, calls an ASR, and performs semantic segmentation. However the SKILL.md claims it uses the OpenClaw built-in LLM for segmentation while the code actually calls an external MiniMax API (api.minimaxi.com). Also the code documents an optional MINIMAX_API_KEY environment variable that is not declared in the skill requirements.
Instruction Scope
SKILL.md instructs running node at a specific workspace path and states segmentation uses the built-in LLM, but the included douyin.js makes outgoing HTTP(S) requests to iesdouyin.com, siliconflow.cn, and api.minimaxi.com and spawns external processes (ffmpeg and curl). The instructions do not declare or warn about the external MiniMax call or the use of curl, giving the agent broader network and command execution scope than the manifest communicates.
Install Mechanism
There is no install spec (instruction-only with an included JS file). That keeps installation footprint small. However the code executes external binaries (ffmpeg and curl) and performs network I/O, so attackers could misuse those execution paths if keys are provided or the code is later modified.
Credentials
The registry metadata requires SILI_FLOW_API_KEY (used for ASR) which is coherent. But the code also reads MINIMAX_API_KEY (optional) and will send text to api.minimaxi.com if present — MINIMAX_API_KEY is not declared in requires.env. The code spawns curl to upload files and post JSON; these elevated network actions are proportional to the stated functionality only if the user expects their audio/text to be sent to those third-party services. The skill does not declare 'curl' as a required binary but uses it, and environment access and network transmission of audio/text to external APIs should be explicitly declared and approved by the user.
Persistence & Privilege
The skill is not always:true and does not request persistent or cross-skill configuration changes. It does write output files under /tmp by default but does not appear to modify other skills or global agent settings.
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-download-v2 - 安装完成后,直接呼叫该 Skill 的名称或使用
/douyin-download-v2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
v2.0.0
Version 2.0.0 introduces major new features and improvements:
- 新增支持从视频中提取语音文案(需 API Key),并自动语义分段
- 支持获取无水印视频下载链接及视频下载
- 增加环境变量 SILI_FLOW_API_KEY 用于调用语音转文字服务
- 提供跳过语义分段的选项 --no-segment
- 更新使用说明与依赖要求
元数据
常见问题
抖音下载+语义分段 是什么?
抖音无水印视频下载和文案提取工具. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 694 次。
如何安装 抖音下载+语义分段?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install douyin-download-v2」即可一键安装,无需额外配置。
抖音下载+语义分段 是免费的吗?
是的,抖音下载+语义分段 完全免费(开源免费),可自由下载、安装和使用。
抖音下载+语义分段 支持哪些平台?
抖音下载+语义分段 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 抖音下载+语义分段?
由 whille(@whille)开发并维护,当前版本 v2.0.0。
推荐 Skills