← 返回 Skills 市场
🔌
ReelOnce-skill
作者
pampas-lab
· GitHub ↗
· v1.0.1
· MIT-0
113
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install reelonce
功能描述
ReelOnce 一体化总控 skill。单次调用即可完成从输入文本到最终视频输出的完整流程：planning、资产图/分镜图/TTS 生成、镜头视频生成、Remotion 工程生成与最终 MP4 渲染。
安全使用建议
Before installing or running this skill: (1) review env.local.example to see which API keys and variables it expects (SKILLS_VIDEO_*, CUSTOM_*, etc.) — the registry metadata omits these; (2) understand that running the skill will install and run the 'reelonce' Python package and npm packages (inspect their provenance and code if you don't trust the source); (3) the pipeline may upload images (URLs or base64) to external video services — don't run it with sensitive local images or credentials present; (4) avoid blindly sourcing env.local from untrusted sources (it can export many secrets); (5) if you want to proceed, run it in an isolated environment (container or VM), inspect the installed reelonce package first, and only provide API keys for services you trust.
功能分析
Type: OpenClaw Skill
Name: reelonce
Version: 1.0.1
The skill bundle is a legitimate wrapper for the ReelOnce text-to-video pipeline. The Python scripts (scripts/agent_api.py and scripts/run.py) safely handle command-line arguments and invoke the underlying package without risky shell execution or obfuscation. The SKILL.md instructions are focused on environment setup and dependency checking, and while they direct users to a specific service (https://skills.video/) for API keys, this is consistent with the tool's stated purpose and lacks any indicators of malicious intent or data exfiltration.
能力评估
Purpose & Capability
The skill is presented as a local orchestrator for a ReelOnce pipeline and the included scripts simply call an installed 'reelonce' package, which is coherent. However, the documentation repeatedly requires/mentions SKILLS_VIDEO_* / CUSTOM_* / COMMERCIAL_* environment variables and instructs users to copy and source env.local — yet the registry metadata lists no required environment variables. That omission is an incoherence: the skill will practically need API keys and environment config to function, but these are not declared.
Instruction Scope
SKILL.md instructs creating a venv, running pip install -e '.[dev]', copying and sourcing env.local, running npm install inside generated Remotion projects, and sending asset references to commercial video services (either URLs or base64-encoded images). These operations are within the stated purpose, but they include reading local files and potentially converting and sending local images as base64 to external endpoints — behavior that should have been made explicit in the metadata and permission model.
Install Mechanism
There is no registry install spec (instruction-only skill with small helper scripts). The README instructs standard pip editable install and npm install in the generated Remotion project. Those are expected for this kind of project and are not themselves suspicious, but they do mean code from the 'reelonce' package and npm packages will be installed and executed on the host.
Credentials
Although the registry lists no required env vars, SKILL.md clearly expects variables like SKILLS_VIDEO_API_KEY, SKILLS_VIDEO_IMAGE_MODEL, SKILLS_VIDEO_VIDEO_MODEL and supports CUSTOM_* / COMMERCIAL_* variants. Requiring API keys and other secrets is proportionate to the task, but the metadata failing to declare them is a mismatch. Also, instructing users to 'set -a && source ./env.local' will export all variables from env.local into the environment — potentially exposing many secrets if env.local contains unrelated credentials.
Persistence & Privilege
The skill is not forced-always (always: false) and does not request system-wide changes in the provided files. It runs subprocesses (calls into the installed 'reelonce' package and will run npm in generated project directories) but does not itself modify other skills or global agent config.
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 reelonce - 安装完成后，直接呼叫该 Skill 的名称或使用
/reelonce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
v1.0.1
- Added --debug-video-prompt option to export the final prompts sent to the video service for each shot.
- Updated usage examples and key parameters to include --debug-video-prompt.
- No other changes to core features or workflows.
v1.0.0
reelonce-skill v1.0.0 - 集成一体化文本到视频总控流程
- 单次调用实现从文本输入到视频输出的完整自动化，包括分镜策划、资产生成、镜头渲染、Remotion 工程生成及最终 MP4 渲染。
- 自动检测并引导用户完成 Python、Node.js 及所需环境依赖与配置。
- 支持自定义项目 ID、镜头数、参考图类型、输出路径等参数配置。
- 明确指引配置文生图与视频生成所需的 skills.video 相关 API Key、模型参数等。
- 优化分工与流程，推荐主流程按 env.local.example 指南配置，兼容旧变量名。
- 输出过程文件、独立数据库、Remotion 工程及最终视频。
元数据
常见问题
ReelOnce-skill 是什么？
ReelOnce 一体化总控 skill。单次调用即可完成从输入文本到最终视频输出的完整流程：planning、资产图/分镜图/TTS 生成、镜头视频生成、Remotion 工程生成与最终 MP4 渲染。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 113 次。
如何安装 ReelOnce-skill？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install reelonce」即可一键安装，无需额外配置。
ReelOnce-skill 是免费的吗？
是的，ReelOnce-skill 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
ReelOnce-skill 支持哪些平台？
ReelOnce-skill 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 ReelOnce-skill？
由 pampas-lab（@pampas-lab）开发并维护，当前版本 v1.0.1。
推荐 Skills