← 返回 Skills 市场
260
总下载
1
收藏
0
当前安装
17
版本数
在 OpenClaw 中安装
/install zenvfx-cli
功能描述
Use this skill when the user needs to create AI videos, manage canvases, nodes, files, or interact with the ZenVFX platform via CLI. Trigger keywords include...
安全使用建议
This skill appears to do what it claims: use the zenvfx CLI to manage ZenVFX canvases and files. Before installing or supplying a token: 1) Verify the CLI package (@tencent/zenvfx-cli) and its publisher on npm/GitHub to ensure you trust the source. 2) Prefer providing a token with the minimal scopes needed (or an ephemeral/token tied to a single project) because the agent can create/delete/upload resources. 3) If you allow autonomous invocation, review what actions you permit the agent to take (or disable autonomous use) so it cannot run destructive commands without your approval. 4) When uploading files, be careful which local paths you expose — the instructions assume you supply local paths for upload. If any of these points are unclear or you want help verifying the npm package or token scope, provide the package URL or your desired risk tolerance and I can guide you further.
功能分析
Type: OpenClaw Skill
Name: zenvfx-cli
Version: 1.0.16
The zenvfx-cli skill bundle provides a comprehensive interface for interacting with the ZenVFX AI video platform. It facilitates canvas management, node operations, and file transfers using a CLI tool installed via a legitimate Tencent npm mirror. The required permissions (ZENVFX_MCP_TOKEN) and capabilities (file uploads to ZenFS/COS) are well-documented and strictly aligned with the stated purpose of AI video generation, with no evidence of malicious intent or prompt injection.
能力评估
Purpose & Capability
The skill declares and uses the zenvfx CLI and a single platform token (ZENVFX_MCP_TOKEN). All required binaries and env vars directly align with the stated purpose of interacting with the ZenVFX platform via CLI.
Instruction Scope
SKILL.md contains concrete shell examples that call the zenvfx CLI to create/open canvases, add/remove nodes, run tasks, and upload local files. That scope matches the described functionality. Note: the instructions include file upload and removal commands (file:upload, file:rm) which — if executed — can read local files specified by the user and modify remote project state; this is expected for a platform CLI but is a capability you should be aware of.
Install Mechanism
The registry entry has no separate install spec, but the SKILL.md metadata suggests installing @tencent/zenvfx-cli via npm from mirrors.tencent.com. Installing from an official vendor namespace (Tencent) is reasonable, but you should verify the package name and publisher before installing. No arbitrary download URLs or archive extraction are present.
Credentials
Only one primary credential is required (ZENVFX_MCP_TOKEN), which is appropriate for a CLI that authenticates to the ZenVFX service. The doc mentions ZENVFX_PROJECT as an optional env var. There are no unrelated secrets or broad system credentials requested.
Persistence & Privilege
The skill is not always-on and has default autonomous invocation allowed. That is typical, but combined with possession of the user's ZENVFX_MCP_TOKEN it would let the agent perform platform actions (create/delete/upload/run) autonomously. This is expected for a remote-control CLI skill, but it raises an operational risk if you grant the token without restrictions.
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 zenvfx-cli - 安装完成后，直接呼叫该 Skill 的名称或使用
/zenvfx-cli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
v1.0.16
No file changes detected in this version.
- No updates or modifications were made; content remains unchanged from the previous release.
v1.0.15
No file changes detected for version 1.0.15.
- No updates or modifications in this release.
- Version remains functionally identical to the previous release.
v1.0.14
- Added project:switch command for one-step project and workspace switching, with automatic matching rules based on userId.
- Workspace automatic matching no longer depends on username, only userId.
- project:switch updates both defaultProject and defaultWorkspace; supports --no-workspace option.
- canvas:node:edit is replaced by canvas:node:props to modify node properties.
- All instructions, command tables, and workflow updated to align with new project and node property commands.
v1.0.13
- Added canvas:node:edit command to support editing node name and position.
- Updated important rules: now canvas:node:edit also triggers auto-save after execution.
- No other command or workflow changes.
v1.0.12
Version 1.0.12
- No file changes detected in this release.
- The functionality, commands, and documentation remain unchanged from the previous version.
v1.0.11
- Added detailed documentation for the file upload command: `file:upload --local-file <本地路径> --file-path <ZenFS路径> --project-id <id>`, including parameter descriptions and usage.
- Expanded the node type table in the "节点参数" section, listing more node types and their application scenarios.
- No changes to code or installation instructions; documentation improvement only.
v1.0.10
No file changes detected. No user-facing changes in this version.
v1.0.9
Version 1.0.9
- Added clarification about auth:login: now documents that most configuration is set automatically, and explains what to do if some fields are not set (saved.defaultProject/defaultWorkspace).
- Added a new command: canvas:node:prompt, for parsing @ image/video placeholders in prompt text in nodes.
- Provided a detailed list of editing commands that auto-save and explicitly advise not to call canvas:save after these operations.
- Improved core usage instructions and error-handling explanation.
- No code changes; documentation enhancements only.
v1.0.8
No file changes detected. No functional or documentation updates in this release.
v1.0.7
- Added a key notice: CLI JSON data is output to stdout only; debug/info logs go to stderr. For correct parsing, filter or ignore stderr when processing output.
- Updated the `canvas:run` command description to clarify that `--wait` now performs built-in polling until completion (with a default 10 min timeout; recommend `--timeout 1800000` for video).
- Enhanced task status query notes: `task:status` output now clearly states that valid output URLs are always in the `url` field, and `download_url` may be empty.
- Updated the core workflow recommendation: now suggests the simpler `canvas:run --wait` method as the default approach for video generation.
- Improved documentation for output handling and clarified recommended polling strategy for long-running tasks.
v1.0.6
No file changes detected for version 1.0.6.
- No code or documentation updates in this release.
- Functionality and interface remain unchanged from the previous version.
v1.0.5
- Added support for new `wsHost` (WebSocket host) configuration in `auth:login` and `config` commands.
- `auth:login` now automatically saves `wsHost` along with existing credentials.
- Updated documentation regarding the treatment of `.canvas` suffix for paths and command usage.
- Minor clarifications and improvements to configuration and workflow explanations.
v1.0.4
Version 1.0.4
- Improved `auth:login`: now auto-saves `defaultProject` and `defaultWorkspace`, eliminating need for `config:set` after login.
- When creating a canvas, `canvas:create` uses `defaultWorkspace` if `--path` is not specified.
- Command guide and sample core workflow updated to reflect automatic project/workspace usage and simpler steps.
- Updated configuration keys: now includes `defaultWorkspace`.
- `auth:login` output now contains enriched metadata including project and workspace details.
v1.0.3
v1.0.3 adds OpenClaw skill installation instructions and node类型/模型查询命令。
- 文档新增“安装 OpenClaw Skill”指引，包含跳过安全扫描提示。
- 新增“节点类型与模型查询”命令：`node:list`、`node:model`。
- 编辑节点类型说明，支持列举和查询模型&枚举值。
- 常用节点类型列表补充：文生图、视频编辑等。
- 其余功能和命令无变动。
v1.0.2
- 新增 `auth:login` 命令，可一键认证并自动保存 token、userId、username。
- `auth:login` 支持可选 `--host`，方便切换不同后端环境。
- 官方文档推荐优先用 `auth:login` 做认证，不再建议手动 `config:set mcpToken`。
- 更新“核心流程”：在所有操作前增加认证步骤，并优化项目设置流程说明。
- 命令速查部分补充 `auth:login` 和认证配置相关说明。
v1.0.1
- Removed all homepage/host and environment switching documentation.
- Updated configuration instructions: only `ZENVFX_MCP_TOKEN` and `ZENVFX_PROJECT` environment variables are referenced.
- Removed host-related installation and config examples to streamline quickstart.
- All command documentation and core workflow remain unchanged.
- Metadata homepage field removed.
v1.0.0
zenvfx-cli 1.0.0
- Initial release of ZenVFX CLI for AI video creation and platform management.
- Supports canvas, node, and file system management via command line.
- Includes commands to create, edit, and run AI video workflows, with session-based canvas operations.
- Full CLI output in JSON format for both success and error cases.
- Installation instructions and core workflows provided in SKILL.md.
- Requires Node.js >= 18 and MCP Token environment/config.
元数据
常见问题
ZenVFX CLI 是什么？
Use this skill when the user needs to create AI videos, manage canvases, nodes, files, or interact with the ZenVFX platform via CLI. Trigger keywords include... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 260 次。
如何安装 ZenVFX CLI？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install zenvfx-cli」即可一键安装，无需额外配置。
ZenVFX CLI 是免费的吗？
是的，ZenVFX CLI 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
ZenVFX CLI 支持哪些平台？
ZenVFX CLI 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（darwin, linux）。
谁开发了 ZenVFX CLI？
由 lenzli（@lenzli）开发并维护，当前版本 v1.0.16。
推荐 Skills