← 返回 Skills 市场
127
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install tekan
功能描述
生成、编辑、协作。一个工具包接入所有主流 AI 模型。只需描述你的创意，即可生成视频、图片和数字人——零手动操作。当用户提到以下任何内容时使用此技能：特看视频、生成视频或图片、数字人、口型同步、文字转语音、TTS、声音克隆、去除背景、商品模特图、图片转视频、文字转视频、AI 图片编辑，或任何创意内容生成工作流——...
安全使用建议
This skill appears internally consistent and implements an OAuth-based client that will: (1) run the provided Python scripts, (2) perform an OAuth device login flow and save credentials to ~/.tekan/credentials.json, and (3) upload files (images/audio) to the service and poll for job completion. Before installing, consider: 1) Verify you trust the service domain (video.tekan.cn) and its privacy/terms because user images and voice samples will be uploaded. 2) Understand that credentials created by the login will be stored on the agent host and grant the skill actions on that account — revoke or delete them if you stop using the skill. 3) Review the bundled scripts if you can (they are included) to confirm no unexpected network endpoints or webhook URLs are used. 4) Be cautious about uploading sensitive personal data (face photos, private voice recordings) or content that might violate rights — voice cloning and avatar creation can implicate likeness/consent. 5) If you run this in a shared or production environment, prefer isolating the agent environment (separate user account or container) so the saved credential file and uploaded artifacts remain under your control.
功能分析
Type: OpenClaw Skill
Name: tekan-video
Version: 1.0.3
The skill bundle is a well-structured integration for the Tekan/Topview AI platform, enabling video, image, and voice generation. It uses a standard OAuth 2.0 Device Flow for authentication, storing credentials locally in `~/.tekan/credentials.json` and interacting with legitimate API endpoints at `api.tekan.cn` and `api.topview.ai`. The scripts (e.g., `auth.py`, `video_gen.py`, `shared/client.py`) handle task submission, polling, and file uploads to S3 in a transparent manner. The detailed instructions in `SKILL.md` are designed to ensure the AI agent correctly manages the asynchronous API lifecycle and provides a consistent user interface, without any evidence of malicious intent or unauthorized data access.
能力评估
Purpose & Capability
Name/description, included Python scripts, and reference docs all match an end-to-end media generation toolkit (image, video, avatar, TTS, background removal, board management). There are no unexpected required environment variables, binaries, or configuration paths that are unrelated to media generation.
Instruction Scope
SKILL.md explicitly requires the agent to execute the bundled Python scripts (auth.py, board.py, video_gen.py, etc.) and to run an OAuth device login flow. That scope is consistent with the skill's purpose, but note: the agent is instructed to run local commands, extract an authorization URL from command output, send a specific-formatted Markdown link to the user, and the scripts will save credentials locally (~/.tekan/credentials.json). The instructions also require persistent polling and automated uploads; these behaviors are expected for this class of tool but give the skill programmatic access to uploaded files and to the saved credentials.
Install Mechanism
No external install spec or remote download is declared; the skill is instruction + bundled Python scripts. Dependencies are installed via a local requirements.txt (pip), which is proportional for a Python-based toolset and does not pull arbitrary code at runtime from unknown URLs.
Credentials
The skill declares no required environment variables or third-party keys. Authentication is handled via an OAuth device flow (auth.py) that yields credentials saved to ~/.tekan/credentials.json. This is proportionate to a web-service client that needs an account on the service; there are no requests for unrelated secrets (AWS, GitHub, etc.).
Persistence & Privilege
The skill stores service credentials locally (~/.tekan/credentials.json) and the agent is required to run the login flow. always:false (not force-included). Because the agent runs the bundled scripts, the skill gains the same privileges as those credentials for the service account; users should be aware of that normal consequence.
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 tekan - 安装完成后，直接呼叫该 Skill 的名称或使用
/tekan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
v1.0.3
No user-visible changes in this version.
- No file changes detected compared to the previous version.
- All features, documentation, and workflows remain unchanged.
v1.0.2
tekan-video 1.0.2
- 新增：微信平台配置说明，登录引导消息需通过 message(action=send) 工具发送，确保 Markdown 链接格式正确传递。
- 登录引导流程规则更严格，明确禁止直接在普通回复中输出 Markdown 链接。
- 其余功能、接口和模块结构无变动。
v1.0.1
## tekan-video 1.0.1 更新日志
- 优化「安装完成与登录引导」：明确要求以 Markdown 链接形式展示授权 URL，禁止裸露链接，提升新手易用性与规范性。
- 增加详细的权限链接插入规则及错误示例，避免因格式问题导致链接失效。
- 视频模型能力描述更新，反映最新可用的先进模型（如 S2.0-白名单版）。
- 调整看板结果链接展示要求，强调需使用 Markdown 链接，提升结果可查阅性。
v1.0.0
tekan-video 1.0.0 – 首个正式版本上线：
- 新增接入特看视频完整 API 工具包，实现图片/视频/数字人/配音/去背景/模特图等全流程自动生成与编辑。
- 完善多模块一站式工作流（avatar4、video_gen、ai_image、remove_bg、product_avatar、text2voice 等）。
- 支持零手动操作：自动上传素材、轮询任务、鉴权、错误处理和编辑结果看板同步。
- 严格认证与权限管理，用户无需手动获取密钥，仅需按照引导链接浏览器登录。
- 全面命令行/脚本控制，所有任务输出都自动集成看板页面链接，方便用户查看和再编辑。
- 提供详细创作与分步操作指南，灵活适应多语言和复杂内容需求。
元数据
常见问题
特看视频 是什么？
生成、编辑、协作。一个工具包接入所有主流 AI 模型。只需描述你的创意，即可生成视频、图片和数字人——零手动操作。当用户提到以下任何内容时使用此技能：特看视频、生成视频或图片、数字人、口型同步、文字转语音、TTS、声音克隆、去除背景、商品模特图、图片转视频、文字转视频、AI 图片编辑，或任何创意内容生成工作流——... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 127 次。
如何安装 特看视频？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install tekan」即可一键安装，无需额外配置。
特看视频 是免费的吗？
是的，特看视频 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
特看视频 支持哪些平台？
特看视频 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 特看视频？
由 特看科技（@wxjhhhh）开发并维护，当前版本 v1.0.3。
推荐 Skills