ai-video-skills

What This Skill Does

The ai-video-skills agent skill is a specialized tool designed for developers and content creators to interact with the skills.video platform via their OpenAPI definitions. It streamlines the lifecycle of generating high-quality AI videos by automating the complexities of REST API requests, SSE (Server-Sent Events) streaming, and asynchronous polling. Instead of manually crafting headers and tracking task IDs, you can rely on this skill to identify the correct model schema, build valid payloads, and handle the connection until a terminal status is returned. By utilizing local scripts to inspect API documentation and OpenAPI specifications, the agent ensures that your request parameters align perfectly with the model's requirements, significantly reducing errors during generation.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/chuyun/ai-video-skills After installation, ensure you have initialized your API key by running the bootstrap verification script provided in the repository. Set your key as an environment variable SKILLS_VIDEO_API_KEY to allow the agent to authorize requests against the skills.video platform.

Use Cases

• Automated Video Generation: Programmatically generate cinematic sequences for video projects without leaving your IDE or command line.
• API Integration Testing: Quickly prototype new video model endpoints by inspecting OpenAPI specs and validating payloads before production deployment.
• Workflow Automation: Integrate video generation directly into custom AI pipelines where you need to transition from text prompts to rendered video files seamlessly.
• Debugging API Calls: Troubleshoot failed generations by using the built-in inspection scripts to check endpoint schemas and requirement constraints.

Example Prompts

1. "Check my current API key and help me generate a video using the kling-v2.6 model with a prompt describing a cyberpunk city under heavy rain."
2. "List all available video generation endpoints from the platform OpenAPI definition and give me a template for the latest text-to-video model."
3. "Run a generation for a 5-second slow-motion shot of a horse running through a field using the /generation/sse/kling-ai/kling-v2.6 endpoint."

Tips & Limitations

• SSE Defaults: Always prefer the SSE-based generation flow as it provides real-time feedback. The agent includes a built-in fallback to REST polling if the stream disconnects unexpectedly.
• Schema Strictness: Do not add fields to your payload that are not explicitly documented in the OpenAPI definition. Use the provided inspection tool to see all required and optional parameters.
• Terminal Status: The agent is designed to wait for a final status (COMPLETED, FAILED, etc.). Avoid interrupting the process while it is in the IN_PROGRESS phase to ensure you receive the final artifact URL.
• Authentication: If you experience 'unauthorized' errors, re-verify your SKILLS_VIDEO_API_KEY environment variable in your shell session.