dsh-vision-assist: Adding Multimodal Recognition Capabilities to Non-Vision Models
The `dsh-vision-assist` plugin intercepts the LLM invocation boundary to address the issue in DeepSeek Harness where pure-text models cannot handle image inputs. When the routed model does not support images, the plugin converts the image into text instructions and triggers the selected multimodal model to perform recognition. It supports hot configuration updates, allowing you to switch models or adjust timeout settings without restarting. The installation command is `dsh plugin --profile web add github:wjling/dsh-vision-assist`, and it requires DSH version >= 0.1.7-alpha.1.
Read MoreDeepSeek Harness Plugin: supanexus/dsh-plugin-chat-enhance
The DeepSeek Harness (DSH) plugin supanexus/dsh-plugin-chat-enhance aims to solve the problem of rapid switching and identification in multi-model services. It introduces a searchable selector, a recent models display, and a multimodal "Image" tab, with support for filtering by name or provider and an Image-only mode. Image uploading reuses the official paperclip feature. Installation requires a full service restart and the use of specific commands. The plugin is standalone, enabling the relevant tab only when a model declares image support, thereby reducing the effort required to operate multimodal models.
Read More