← 返回 Skills 市场
72
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install video-object-ops
功能描述
Skip the learning curve of professional editing software. Describe what you want — remove the background person and replace it with a solid color — and get o...
安全使用建议
This skill appears to do what it says: it uploads videos to a remote render service and returns edited clips. Before installing: 1) Confirm you trust the external host (https://mega-api-prod.nemovideo.ai) — videos and possibly sensitive visual content will be uploaded to their servers. 2) Prefer using the anonymous-token flow (ephemeral token) if you don't want to store a permanent NEMO_TOKEN in your environment. 3) Note the SKILL.md frontmatter references a local config path (~/.config/nemovideo/) but the registry metadata omitted it — ask the publisher whether the skill will read/write that directory. 4) If you must keep private footage local, do not use this skill; or check the service's privacy/retention policy. 5) If you want higher assurance, request a verified source/homepage or inspect a published package/release for the service before enabling the skill.
功能分析
Type: OpenClaw Skill
Name: video-object-ops
Version: 1.0.0
The video-object-ops skill is a legitimate integration for a remote video editing service hosted at nemovideo.ai. It provides detailed instructions for an AI agent to manage authentication, session state, file uploads, and cloud-based video rendering. The skill includes security-conscious directives, such as preventing the agent from displaying raw tokens or JSON to the user, and its data access is limited to its own configuration directory (~/.config/nemovideo/) and the NEMO_TOKEN environment variable.
能力评估
Purpose & Capability
Name/description (video object editing) align with the required credential (NEMO_TOKEN) and the SKILL.md instructions (upload video, create session, render). The requested environment variable is proportional to the stated purpose.
Instruction Scope
Instructions focus on authenticating, creating a session, uploading videos, sending SSE edit messages, and polling render status — all within the service domain. They explicitly tell the agent not to print tokens or raw JSON. They also instruct deriving attribution headers and detecting an install path to set X-Skill-Platform, which may require checking local paths; this is not strictly needed for core functionality but is low-risk. The SKILL.md frontmatter references a config path (~/.config/nemovideo/) which would imply reading/writing local config, but the registry metadata listed no required config paths — a small inconsistency.
Install Mechanism
Instruction-only skill with no install spec and no code files. Lowest-install risk; nothing is downloaded or written by an installer.
Credentials
Only NEMO_TOKEN is declared as required and is the primary credential, which matches the need to call the remote API. The skill also describes an anonymous-token flow (generate UUID and request a 7-day token) so a token can be obtained without pre-provisioning credentials. The SKILL.md frontmatter mentions a config path (~/.config/nemovideo/) which could permit local token storage — this is related to the skill's purpose but the registry metadata omitted it, creating an inconsistency to verify.
Persistence & Privilege
always is false and model invocation is not disabled (default). The skill does not request persistent system-wide privileges or modify other skills. Autonomous invocation is allowed by default; that is normal but increases blast radius if a malicious skill were present (no evidence of that here).
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 video-object-ops - 安装完成后，直接呼叫该 Skill 的名称或使用
/video-object-ops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
Initial release of Video Object Ops — AI-powered video object editing.
- Upload MP4, MOV, AVI, or WebM files (up to 500MB) for fast, frame-accurate object edits.
- Describe edits in plain language (e.g., "remove the background person and replace it with a solid color") and receive object-edited clips in 30–90 seconds.
- Automatic setup and session management with free anonymous tokens; no need to install anything.
- Supports common export formats (MP4, MOV, AVI, WebM, etc.) and easy download links.
- Ideal for creators looking to manipulate video objects without advanced editing software.
元数据
常见问题
Video Object Ops 是什么？
Skip the learning curve of professional editing software. Describe what you want — remove the background person and replace it with a solid color — and get o... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 72 次。
如何安装 Video Object Ops？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install video-object-ops」即可一键安装，无需额外配置。
Video Object Ops 是免费的吗？
是的，Video Object Ops 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Video Object Ops 支持哪些平台？
Video Object Ops 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 Video Object Ops？
由 mory128（@mory128）开发并维护，当前版本 v1.0.0。
推荐 Skills