Vision tools · DeepSeek Harness plugins — Page 12

Custom background image plugin for the DeepSeek Harness web GUI: upload a local picture or paste an image URL, and render it behind the whole app surface with adjustable opacity, scrim, panel transparency and frosted-glass blur.

Out-of-tree dsh provider plugin: a DeepSeek gateway route that claims image input and transparently describes pasted images through a configured vision-language model (e.g. Qwen-VL) before the text-only DeepSeek wire sees them.

DSH plugin: start/stop a local llama.cpp llama-server from the settings → plugins page — two model slots (A/B) with per-slot model-file selection, eight editable launch-parameter groups (text/vision × fast/long × 2 slots), and one shared DSH provider carrying both slots' models with four thinking-effort levels.

AX-first macOS computer use for DeepSeek Harness: accessibility snapshots with stable refs, background input, Unicode typing, menus, apps, set-of-mark screenshots, on-device OCR, verified actions, and optional Jev (TypeSafe System One) semantic guards — 11 tools bridged to the macos-cu CLI.

dsh-mineru: MinerU 文档解析插件 for DeepSeek Harness — 多模态全格式 (PDF/Word/PPT/Excel/HTML/图片) → 结构化 Markdown. 填 Token 走精准解析 API, 留空走 Agent 轻量解析 API.

DSH注入系统提示词和设置注入上下文功能插件，插件内置鲸鱼娘社区人设！本文中和插件设置界面都将注入系统提示词的行为称作“人设正文”，将设置注入上下文的行为称作“历史对话”。可通过在设置界面人设管理保存为预设“人设”，并且支持将保存的人设正文和历史对话作为数据块选择PNG图片或作为JSON文件导出，用PNG图片本身作为封面。注入的系统提示词可以自行编辑注入位置。支持屏蔽dsh本身注入的系统提示词、runtime上下文、schema工具以及其他user侧内容，使得发送内容保证让模型仅接收对话内容，且拥有工具保活通道，避免后续调用工具仅输出文本。历史对话可通过DSH本身导出(/export)的Session日志直接导入，用仿照DSH本身注入标签的格式包裹。各会话单独保存设置和人设，人设切换和设置勾选项不影响新会话和进行中的会话。

dsh-plugin-appshot — a DeepSeek Harness plugin for macOS and Windows context screenshot capture.

QQ ↔ DeepSeek Harness agent bridge over OneBot v11 (reverse WebSocket). v0.5.x "See it, find it": merged-forward chat records are expanded for the model instead of being dropped, group member/roster and group-info queries, recent-history and emoji-reaction agent tools, group file listing and download, group albums, image OCR, and a keyword-searchable per-day message archive (/找) — plus the whole v0.4 debugging stack (trace id on every message, a Chinese reason on every silent drop, structured JSONL events, live SSE event stream with per-message decision chains, one-click diagnosis, diagnostic-bundle export, offline sandboxed dry-run replay and event injection, a 6-hard-constraint acceptance page) and a standalone control console (control/). Independent bundle; remove anytime: dsh plugin --profile web remove dsh-qq-onebot-bridge

DeepSeek Harness 桌面端点选与圈选修改：结构化定位、关键样式和原文件引用

see_image tool for DSH: route image files to a configurable vision model (OpenAI-compatible API) and relay its description back to a text-only model

Turbo Vision-style character-cell windowing IDE for DeepSeek Harness agents

OMX 风格工作流技能集（oh-my-deepseek-harness）：移植自 oh-my-codex（MIT）的 deep-interview / ralplan / ralph / plan / autopilot / team / code-review 等技能，全部重写为 DeepSeek Harness 原生机制（goal / workflow / ralph / subagent / ask_user_question / read_image 工具），含视觉技能 visual-ralph。

Image preview in the dsh session: read_image results render as a thumbnail and open a full-size lightbox with zoom, rotate, pan, download, and cross-row navigation

A DeepSeek Harness Plugin for Previewing Mermaid Diagram by Images in Chat Sessions. View on GitHub (https://github.com/baconbao/dsh-mermaid-image-preview).

DeepSeek Harness Computer Use 插件（Windows）：让 AI 看屏幕、点鼠标、敲键盘操控电脑；控制时屏幕泛起橙色光晕并显示“DeepSeek Harness 正在操控你的电脑”，按 Esc 暂停 / 继续。

Dynamic attachment presentation plugin for conversation input, message-image, and trajectory image slots

DeepSeek Harness plugin: reasoning-effort auto-fill declarations and progressive Figma MCP bridging (design-to-code + AI-driven design).

Feishu (Lark) entry point for DSH: one-command install plugin that lets users drive DSH agents from a Feishu bot over the long-connection event channel, with image/file/video/audio attachments, interactive ask/approval cards, live streaming progress cards with a tool panel, and proactive feishu_send pushes.

DSH 视频证据插件：带时间戳画面、字幕/ASR 与 OCR 原生返回当前视觉模型；按需细看、缓存和可选独立 API

QQ Bot channel bridge for DeepSeek Harness

File explorer for the DSH dock: browse the active conversation's working directory and open files in the registered file viewers.

Automatic Web image-to-text bridge plus vision and OCR tools for DeepSeek Harness

Codex-backed web search, image generation, and image understanding tools for the DeepSeek Harness.

Multi-agent collaboration suite for DeepSeek Harness: a user-configured specialist roster with on-demand dispatch (team_call / roundtable), model comparison, and a multimodal vision bridge — models come from the official provider flow, no bundled adapters.