dsh-chat-imagine
@corrinehu
Multimodal
Automatically calls image generation tools (API or local CLI: mmx/codex/agy) in DSH chat and displays images; also supports image recognition via CLI vision capabilities.
All DeepSeek Harness plugins
49 pluginsGitHub data updated Aug 23, 2026
dsh-promotion-toolkit
@lhmd
Multimodal
Turns any idea into platform-native promotional content for 17 platforms, with structured output and platform-specific adaptation.
dsh-image2-draw
@JuneLearn
Multimodal
Image2 generation plugin for DeepSeek Harness via third-party OpenAI Images-compatible APIs: configure baseURL and API key, use gpt-image-2 for text-to-image and image-to-image.
novel-writer
@sailoumili
Multimodal
A multi-agent novel writing mode with a coordinator and five specialized sub-agents (world, plot, character, prose, review) that collaboratively write Chinese novels with structured workflows and file-based project management.
dsh-mmx-bridge
@welsione
Multimodal
A single tool that bridges DSH text-only models to MiniMax's full multimodal stack: image understanding, image/video generation, speech synthesis, music generation, audio cover, web search, and quota queries.
wanjiqi-meme
@Chu-Xin-r
Multimodal
An AI Skill containing 22,771 real danmaku memes from the 6657 live stream, enabling generation of streamer-style comments, commentary, and CS/DOTA crossover jokes.
dsh-ears
@WizisCool
Multimodal
Voice input plugin for DeepSeek Harness supporting multiple ASR backends (Web Speech, local Whisper, Groq, DashScope, OpenAI-compatible) and LLM-based polishing.
dsh-office-tools
@kw78
Multimodal
Model-facing Office tools for DeepSeek Harness to create, read, and update Word, Excel, and PowerPoint files with workspace-safe paths and PPT image embedding.
dsh-vision
@54xkeee
Multimodal
Vision plugin for DeepSeek Harness providing image recognition via Doubao Web (zero-cost, no API key), Antigravity IDE quota, Gemini, and other channels with auto detail escalation and evidence memory.
pptpress
@liustack
Multimodal
Stable, editable PPTX generation for AI agents using semantic IR and native DrawingML output. Includes DSH plugin, Claude Code plugin, and CLI.
dsh-director-toolkit
@lhmd
Multimodal
Turns rough 3D ideas into direction packs for Blender, Three.js, Houdini, or C4D, including camera, lighting, materials, render settings, build steps, and bilingual showcase copy.
dsh-novel-writer
@siweina
Multimodal
Novel writing assistant plugin for DeepSeek Harness: chapter library management, sentence pattern analysis (nine types, emotion curves, style fingerprint), style self-check, plot tracking, batch import, and AI-assisted continuation writing.