DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

Model Auto Enricher — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
M

dsh-model-auto-enricher

Model Auto Enricher

DSH Model Completion Assistant: eliminates the tedious process of manually opening dozens of models one by one after adding a new model channel in DSH to adjust context windows, enable multimodal image input, and configure reasoning effort tiers (low/high/max). Supports automatic completion upon sav

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:66rapper/dsh-model-auto-enricher#4c18dc2a231f336a31d49dccfff9606edc6456c9
READMECompatibilityVersions

Description

DSH Model Completion Assistant: eliminates the tedious process of manually opening dozens of models one by one after adding a new model channel in DSH to adjust context windows, enable multimodal image input, and configure reasoning effort tiers (low/high/max). Supports automatic completion upon saving newly added models and one-click batch completion, with automatic timestamped backups before writing. Model channel auto-enricher for DeepSeek Harness — eliminates manually configuring context windows, image input and reasoning effort tiers model by model.

Compatibility and provenance

Model Auto Enricher is published as dsh-model-auto-enricher and currently resolves to version 2.0.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
web
Release source
github
Registry updated
10/19/2026

Versions

2.0.0stable
10/19/2026

Related plugins

Loading related plugins…

Latest
2.0.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
web
License
MIT
Source
github
GitHub
★ 0
Weekly downloads
0
Last push
10/19/2026
View source ↗
README badge

Click the badge to copy Markdown for your README.

Do you maintain this Plugin?Claim benefit · Priority security scan

Verify the GitHub repository declared in package.json to manage this listing. After you claim it, Hub will prioritize a security scan of the current version and publish the result when it passes.

Claim this Plugin →
Report an issue

Related plugins

More verified plugins in models-usage.

Cost Meterdsh-cost-meterDeepSeek Harness session cost statistics plugin: current session cost, daily expenses, historical records, and synchronization with official pricing; supports multi-provider and multi-model price billing (built-in price catalog for 170+ model IDs with automatic matching), and querying and displayingCodex Subscriptiondsh-codex-subscriptionUse ChatGPT and Codex subscriptions in DeepSeek Harness with OAuth, quota, safe resets, web search, images and Fast mode, and control DSH from the ChatGPT mobile app (Beta)Connect Workbuddydsh-connect-workbuddyConnect locally authenticated WorkBuddy models to DeepSeek Harness with read-only credit overview and model management.Connect Traedsh-connect-traeConnect locally logged-in Trae models to DeepSeek Harness: run domestic and international providers in parallel, with usage/points overview and daily check-in rewards.

README

⚡ DSH 模型补齐助手(dsh-model-auto-enricher)

告别在 DSH 里加完模型渠道后,还要把几十个模型一个个手动点开调上下文、勾图片、配思考强度的繁琐苦力活!


😫 为什么做这个插件?(真实痛点)

在 DSH 里每添加一个新的中转站或第三方 API 渠道,里面往往有十几个甚至几十个模型。 但是在 DSH 默认设置界面里,拉取或添加完模型后,每一个模型都得你自己手动点开去调:

  • 要一个个手填 上下文窗口(默认是空的,或者被限制在 128k / 200k,不改很快就会爆上下文);
  • 要一个个手填 最大输出 Token(默认是空的);
  • 要一个个手动勾选 「图片」多模态开关(默认只勾了纯文本,不勾根本发不了截图);
  • 要一个个手动打开 「推理等级」开关 并添加 max 档位(默认没开,开了也只有 low / medium / high,没有最高 max 档)。

十几个模型一个个点开改下来,又麻烦又累;以后每次往渠道里新加一两个模型,还得再重复折腾一遍!

装上 「DSH 模型补齐助手」 后: 👉 你只需要在设置里把渠道和模型加进去、点一下保存,剩下的调上下文(100 万)、开多模态发图、配思考强度档位(low / high / max)全部由插件在后台自动帮你一次性搞定!


🛠️ 具体帮你省掉哪些手动操作?

你以前要手动挨个调的项DSH 默认刚加进来时装上本插件后(全自动)
1. 调上下文容量空白未填,或被截断在 128k / 200k / 256k自动全部拉满到 100 万长上下文(1,000,000)
2. 调最大输出空白未填自动按系列补齐 64,000 / 128,000 最大输出
3. 调多模态(发图)默认只勾了「文本」,聊天框发不了图自动为所有模型勾选 「文本 + 图片」([text, image])
4. 调思考强度档位默认关闭,或手动打开后缺少 max 档自动配齐 { low, high, max } 思考强度档位(并保留已有自定义档位)

✨ 核心亮点(v2.0 中文友好 · 全自动版)

  1. ⚡ 添加新模型保存即自动生效(无需手动点按钮、无需重启 DSH)
    • 插件在后台自动监听 DSH 的模型设置保存事件。
    • 只要你在 「设置 → 模型」 里添加了新模型并点击「应用」,插件会在 0.6 秒内自动在后台帮你把百万上下文、64k+ 最大输出、发图支持、思考档位全部补齐,并立即热重载生效!
  2. 🧠 认识所有新模型(内置全系列规则 + 通用新模型兜底)
    • 不仅完整识别 GPT-4o / GPT-5 / GPT-6 / o1~o9 / Codex、Claude 4 / 4.5 / 4.6 / 5 / 5.5+ (Opus / Sonnet)、Gemini 2.5 / 3 / 3.5 / 4、DeepSeek V3 / V4 / V4.1 / R1 / R2、GLM 5 / 5.2 / 5.3 / Z1、Kimi K2.6 / K2.7 / K2.8 / K3、腾讯混元 Hy3 / Hy4、MiniMax / Step / Qwen / Grok / 豆包 / Llama / Mistral 等全系模型;
    • 哪怕是你以后新加的、名字从来没见过的全新模型或内部代号模型(如 space-bunny、argon、astra、sol 等),也会自动触发**「通用新模型规则」**,默认按顶配规格(1M 上下文 + 64k 最大输出 + 支持发图 + low/high/max 思考档位)一步到位补齐!
  3. 🛡️ 100% 安全,只升不降,写前自动备份
    • 只升不降:如果你某个模型已经设了更大的上下文或更大的输出(如 128000),绝不会给你改小;已有的自定义思考档位(如 off、medium、xhigh)一律原样保留。
    • 自动备份:每次修改配置文件前,都会自动在同目录生成带时间戳的备份文件(cordis.patch.yml.bak-auto-enrich-YYYYMMDD-HHmmss)。
    • 严格幂等:所有模型已经是最佳状态时,不会改动任何字节,也不会产生多余备份。
  4. 🖥️ 内置可视化管理面板(设置 → 模型补齐助手)
    • 在 DSH 设置页左侧导航栏新增 「模型补齐助手」 面板,随时查看所有渠道与模型的健康状态、识别系列标签、上下文容量、多模态与思考档位明细,也支持手动一键补齐。

📦 一键安装

方式一:在 DSH 插件市场(dshmarket)中安装

打开 DSH 设置 → 插件市场,搜索 dsh-model-auto-enricher 或 模型补齐助手,点击安装即可。

方式二:命令行一键安装

桌面版(desktop profile):

dsh plugin --profile desktop add 66rapper/dsh-model-auto-enricher

网页版(web profile):

dsh plugin --profile web add 66rapper/dsh-model-auto-enricher

📊 内置智能补齐规则表(v2.0)

模型系列自动识别关键词补齐上下文 (contextWindow)补齐最大输出 (maxTokens)模态支持 (input)思考强度档位 (reasoningEfforts)
Gemini 系列gemini / gemma1,000,000 (1M)64,000[text, image]除纯 lite / image 生图小模型外全系开启
Claude 系列claude1,000,000 (老款 Claude 2/3 为 200,000)64,000[text, image]4.5+ / 5+ / opus / sonnet / thinking / high 全系开启
DeepSeek 系列deepseek1,000,000 (1M)128,000[text, image]v4 / v5 / r1 / r2 / pro / flash / thinking 全系开启
GLM 智谱系列glm / chatglm1,000,000 (1M)flash 128,000;其余 64,000[text, image]5.2 / 5.3 / 6 / z1 / flash / high 开启
Kimi 系列kimi / moonshotk2.6 / k2.7 为 256,000;其余 1,000,000k2.8 / k3+ 为 64,000;其余 32,000[text, image]k2.8 / k3 / k4 / preview 开启
GPT 系列gpt-* / chatgpt / o1~o9 / codex1,000,000 (1M)64,000[text, image]gpt-5 / 6 / 7+ / o1~o9 / codex / astra / sol 开启
腾讯混元系列

📄 开源协议

MIT License

hy3 / hy4 / hunyuan
hy3 为 192,000;hy4+ 为 1,000,000
64,000
[text, image]
hy3 / hy4+ / preview / thinking 开启
主流大模型系列minimax / step / qwen / grok / doubao / llama / mistral1,000,000 (1M)64,000[text, image]step-5+ / grok-3+ / qwen-3+ / m3+ / thinking 开启
🌟 通用新模型规则任何未见过的全新模型 / 内部代号模型(自动兜底)1,000,000 (1M)64,000[text, image]默认开启 { low, high, max }
GPT 开源版gpt-oss128,00064,000[text](纯文本)不开启
残缺档位自动补全任何已开启部分思考档位的模型保持原值或拉升保持原值或拉升保持或补齐图片自动补齐缺失的 { low, high, max },保留原有档位