dsh-model-vision：自动补全 DSH 模型清单里漏掉的「可看图」能力
DSH 通过「获取可用模型」同步清单时往往只记录名称与上下文，漏掉网关已公布的图片输入能力，导致贴图被拒或手动勾选后被覆盖。dsh-model-vision 是 charovo0059 维护的 MIT 插件，优先读网关声明，必要时用极小图探测，只写「输入类型：图片」与「· 可看图」标注。安装命令为 dsh plugin --profile web add dsh-model-vision，装后需重启；支持定时对账与设置页一键检测。适合 OpenAI 兼容网关与自建服务，安装前应审阅源码并以 DSH 进程权限理解其读写范围。
Read Moredsh-bundle-vision: Zero-Core-Change DeepSeek Harness Vision Capability
`dsh-bundle-vision` is a plugin package that provides vision capabilities for DeepSeek Harness (DSH). Through the Bundle patch mechanism, it mounts visual functionality to any Profile without modifying core code. The plugin registers the `describe_image` tool, which supports reading local images and invoking an external multimodal routing service. It relies on the file system, attachment service, and LLM calls, enabling pure text models to process images flexibly for agent development scenarios that require visual capabilities.
Read Moredsh-design-qa: Enabling Image Understanding for Text-Only Models in DeepSeek Harness
The dsh-design-qa plugin enables text-only models to “read” images by integrating multimodal tools, allowing them to perform visual judgments within the product design loop. It supports on-demand invocation, does not consume the main model's context, and includes built-in evaluation benchmarks. Installation requires configuring an Alibaba Cloud Bailian API Key, and it defaults to using the qwen3.8-max model. The plugin emphasizes that “visibility does not equal judgment capability” and should be used according to specific scenarios.
Read Moredsh-deepseek-vision: Adding image recognition to DeepSeek text models
dsh-deepseek-vision is an MIT-licensed DSH plugin designed to provide image recognition capabilities for text-only DeepSeek models. By registering the `deepseek_vision` tool, the plugin allows the model to submit local image paths (supporting PNG, JPG, WebP, and GIF, but not HEIC) and optional prompts. When invoked, the plugin utilizes the DSH subprocess service to safely call the local `deepseek-vision-cli` via array parameters, sends the image to the DeepSeek Web endpoint for recognition, and returns the descriptive text to the model. Users need to establish a Web session via manual login to the assistant or Token login; the Token is stored only in local cache. Installation supports local directories or Releases.
Read Moredsh-plugin-deepeye: A Cordis Plugin Providing Visual Capabilities for DSH
`dsh-plugin-deepeye` is a native Cordis vision plugin designed for DeepSeek Harness (DSH), aimed at resolving the tedious issue of manually invoking external vision services. Developed by Favio8 under the MIT license, it supports Node >=22.19.0 environments. Its core features include tool-based invocation capabilities such as image description, OCR extraction, visual question answering, UI layout analysis, and clipboard screenshot analysis. The plugin supports multiple backends, including OpenAI, Google Gemini, and custom OpenAI-compatible services, and offers functionalities like LRU result caching, automatic image scaling preprocessing to save tokens, and non-blocking asynchronous reading of the system clipboard. Additionally, it supports use in plain-text sessions through
Read More