thinkingpeach-sketch/wan3-agent-skills

thinkingpeach-sketch★ 0JavaScript最后同步: 2026-08-19

在 GitHub 打开

Portable WAN3 image and video generation skills for AI coding agents

README 摘要

WAN3 Agent Skills Portable image and video generation skills for agents that can run local commands. They call the WAN3 Developer API, use the caller's WAN3 account credits, poll asynchronous jobs, and download completed media. This repository contains: - skills/wan3-image-generation - image generation and editing - skills/wan3-video-generation - text, image, reference, and video workflows - integrations/deepseek-harness - optional DeepSeek Harness discovery adapter Requirements - Node.js 20 or newer - A WAN3 account with credits - A user-created WAN3 API key from The skills do not require or contain an upstream provider credential. Install Copy the two skill folders into the skill directory used by your agent. For Codex: For Claude Code, copy them into /.claude/skills/ . For DeepSeek Harness, copy them into the project .dsh/skills/ directory or the user DSH skill directory; the optional adapter has additional instructions in integrations/deepseek-harness/README.md . Keep the API key in a local secret manager or environment variable: Do not commit the key, paste it into an issue, or place it in browser code. Verify the connection The clients discover the current model and input con…

在 GitHub 查看完整 README →
Agent/智能体ai-agentscodexdshdsh-pluginimage-generationvideo-generationwan3agent

分类