Plugins

On-demand vision for text-only DSH sessions: images become markers, and a vision_describe tool sends only image + question to an OpenAI-compatible vision model

Auditable vision and cross-platform Computer Use runtime for DeepSeek Harness with source-preserving evidence.

Bridges session images to configurable vision providers and returns text-only analysis to eligible DeepSeek Harness model routes.

Native LLM-provider vision bridge: images pasted in the chat are described by a vision model (Qwen3-VL via pi-ai/llama.cpp) and the text description is fed to text-only DeepSeek for the reply — image admission, routing and compaction all run through harness-native mechanisms, with an LRU description cache and 503 retry.

Composer-attached images are transcribed to text by an OpenAI-compatible vision model before reaching text-only DeepSeek models.

DSH 双模型路由插件：主对话留在 DeepSeek，图片任务自动分发到显式配置的视觉模型，再把分析结果交回 DeepSeek 继续。

明眸 VisionBridge - 自研视觉桥：瞎子模型收图时自动调用视觉模型识别，把识别文字喂回主模型，无感、可配置、升级不丢

Self-owned vision bridge for DeepSeek Harness: pick the vision model yourself from the available ones. Image content blocks are rewritten into text markers at the agent boundary so text-only models never fail a turn, and a describe_image tool answers via