The "visual-primitives" topic on GitHub groups 1 open-source project in the DeepSeek Harness (DSH) ecosystem, led by dsh-tool-visual-primitives with 1 GitHub stars. dsh-tool-visual-primitives — DeepSeek Harness vision enhancement plugin: hands images to an external vision model for analysis and outputs plain-text evidence with coordinate-based visual primitives, so non-multimodal text models can also understand images, screenshots and documents in conversation. Every project here is indexed by DSH Universe with live GitHub data — stars, activity and install status — so you can compare and install directly.
DeepSeek Harness vision enhancement plugin: hands images to an external vision model for analysis and outputs plain-text evidence with coordinate-based visual primitives, so non-multimodal text models can also understand images, screenshots and documents in conversation.