Routes images to a vision model and feeds the textual description back to the text-only DeepSeek agent.
dsh plugin --profile web add dsh-vision
View the source on GitHub