Adds multimodal vision capabilities from DeepSeek's official model to agents, enabling image understanding across multiple coding tools.
dsh plugin --profile web add github:xlight/deepseek-visionary#path:packages/dsh-plugin
View the source on GitHub