modlens
@liustack
Multimodal
The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics).
All DeepSeek Harness plugins
25 pluginsGitHub data updated Aug 23, 2026
dsh-vision-router
@ysr666
Multimodal
dsh-vision-router provides eyes for text-only DeepSeek Harness agents with a built-in free vision chain and pixel-level vision tools including Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, and screenshots.
dsh-vision-toolkit
@Anionex
Multimodal
dsh-vision-toolkit is a vision toolbox for text-only DeepSeek Harness models, offering image Q&A, long-screenshot OCR, UI restoration, grounding, pixel diff, Artifacts, and Web UI integration.
superdesign-skill
@superdesigndev
Multimodal
A design skill for coding agents that provides design judgment, generates and iterates UI, presentations, and graphics on an infinite canvas, with cross-session continuity.
dsh-openpencil
@ZSeven-W
Multimodal
The DeepSeek Harness plugin for OpenPencil — preview, inspect, and edit real .op documents inside a conversation.
dsh-image-gen
@shanliuling
Multimodal
Generates images directly in DeepSeek Harness chats. Supports multiple providers including Google Gemini, OpenAI Images, OpenAI Compatible API, and ByteDance Seedream/Volcengine Ark. Includes a gallery tab for viewing and searching history, full-screen preview, copy, and download.
dsh-oil-creator
@oil-oil
Multimodal
AI-assisted local creator workbench for DeepSeek Harness, organizing content projects as folders.
dsh-vision
@oil-oil
Multimodal
Near-native image understanding for DeepSeek Harness, with vision bridge for text-only models.
dsh-univer-office
@dream-num
Multimodal
Preview, create, edit office spreadsheets, docs & slides inside DeepSeek Harness, powered by Univer.
dsh-meme
@yyh-001
Multimodal
Meme plugin for DeepSeek Harness — find, send, and learn memes; text-only meme battles, emotion-triggered sending, QQ/WeChat-style sending, AI auto-learning, custom meme packs, and multiple styles.
dsh-omi-voice
@PolinniZhong
Multimodal
DeepSeek Harness plugin for reading AI replies aloud with natural Doubao TTS voices, featuring BYOK, click-to-read, pause/resume, and local-only processing.
picturereader
@jing-hy
Multimodal
DSH plugin providing pixel-to-text image reading for text-only models, with image_scan/image_ocr/image_sample tools and an image-reading skill, fully local with optional PaddleOCR.