kaixinbaba/dsh-vision-recognizer ↗★ 0
dsh-vision-recognizer
DeepSeek Harness 图像识别插件。保持 DeepSeek 作为对话大脑，同时调用可配置的第三方视觉模型（支持 15+ 供应商）来转录和解析附加的图片。
AI 分析
核心用途是为 DSH 补充图像识别能力。适合希望在使用 DeepSeek 进行核心推理的同时，利用其他视觉模型（如 OpenAI 兼容或 Anthropic 接口）解析图片输入的用户。
安装
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:kaixinbaba/dsh-vision-recognizer说明文档
阅读完整 README ↗Configuration storage
Config saved from the UI is written to $DSH_HOME/vision-recognizer.json and merged over the bundle defaults at startup. cordis.patch.yml only carries factory defaults; a user cordis.patch.yml override still works as the composition-time fallback.
⚠️ patch semantics: the bundle's
- insert:appends this row to the entry list. Writing a second- insert:with the same id in your owncordis.patch.ymlwould register the adapter twice (undefined behavior). To override individual keys, write a single top-level- id: dsh-vision-recognizerentry; better yet, use the Settings UI.