DeepSeek Harness 플러그인 — 742 페이지

Pin-and-comment annotation for the DeepSeek Harness image lightbox: drop numbered pins, type a note, and the original image plus a coordinate list lands in the composer. 图片图钉标注：在图片上钉图钉写评论，自动生成带坐标的说明一起发出去。

DSH 会话内图片预览服务：把工作区图片通过 HTTP 暴露给会话渲染、文本→图片渲染（双字体防乱码）、SVG→PNG 转换。AI 在回复里贴 markdown 图片 URL，用户直接在对话中看图。

Inline previews and clipboard copying for DeepSeek Harness read_image tool calls, including nested Code Mode calls.

Local image hosting for DeepSeek Harness: serves images from a preview directory over the web profile's own HTTP server so chat messages can render them inline, plus a preview_image tool that stages an existing local image and returns its URL.

Inline previews for the images an agent already produced in DeepSeek Harness: read_image attachments and local paths in any format (HEIC/TIFF/PSD/SVG/BMP/ICO/TGA/PNM/JP2 …), rendered through the official gallery and lightbox.

在 dsh Web 对话正文中直接渲染本地图片路径的插件（host 改写路径 + 极简客户端拦截点击）

DeepSeek Harness plugin: structured image analysis via vision APIs (read_image_mimo tool) with Web UI config card. Attachment input moved to dsh-input-enhancement.

Give DeepSeek Harness agents the ability to read images directly: a model-facing read_image tool that answers questions about an image through any OpenAI-compatible vision endpoint.

DeepSeek Harness plugin: analyse images out of band with a vision model — pasted images are digested into text before admission and a describe_image tool covers image paths, all without ever changing the session's model.

User-tunable inline image display for dsh web sessions: a larger, auto-unrolling read_image tool view behind an image-settings section.

An Images workspace for DeepSeek Harness: one page to generate with fal.ai, browse every result, compare candidates, and animate the winner with Kling image-to-video.

DeepSeek Harness plugin: an inline conversation image card with a bottom-left edit button (annotate / erase / resize) that hands the edit back to the agent.

Shows oversized images to the model at a readably high resolution by tiling them before admission.

DSH plugin that lets text-only (non-multimodal) models send and read images: bypasses the image admission gate, rewrites image blocks to local paths, and OCRs/describes them via configurable vision API providers (image_to_text tool).

图片处理：读取图片尺寸/格式、缩放与格式转换（内置 ffmpeg，跨平台）

DeepSeek Harness 原生插件：让纯文本模型（如 DeepSeek）经 agent/pre-step 劫持 + resolveModelInfo 包装自动识别上传图片（qwen-vl），使纯文本对话也能“看见”图片。

A proper image viewer for DeepSeek Harness: zoom, pan, rotate, download, copy and gallery browsing for every image the chat renders.

Seamless image understanding for DeepSeek Harness: lets pure-text models read pasted/dropped images and Feishu/Lark images via a configurable vision model, without touching the host.

Re-encode WebP attachment request-images to PNG for LLM backends (e.g. llama.cpp) that cannot decode WebP. Fixes the '400: Failed to load image or audio file' error for alpha-channel images.

Image2 (gpt-image-2) generation plugin for DeepSeek Harness with simple relay settings, in-conversation previews, a chat-dock studio (text-to-image / image-to-image / multi-view character consistency) and a reusable prompt/flow preset library. Community fork of JuneLearn/dsh-image2-draw with gateway compatibility fixes.

DeepSeek Harness bundle that exposes an OpenAI-compatible image relay as the image2_generate tool for text-to-image and image-to-image, with inline media previews.

image_gen and image_vision model tools for the DeepSeek Harness, powered by the ChatGPT subscription (no API key).

Local image editing toolkit for DeepSeek Harness: one image_edit tool for deterministic edits — cutout (quick flood-fill or rembg AI), trim, flip, rotate, brightness/contrast/saturation, blur, sharpen, rounded corners, border, canvas resize+center, sprite sheets, and PNG/JPEG/WebP export. Also ships as a skill (SKILL.md + CLI scripts).

生成图片：通过 New API 中转站（guaihub 等）调用 OpenAI 风格 /v1/images/generations，并把图片直接显示在 DSH 对话里。