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.

DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
G

dsh-gateway-models

Gateway Models

DeepSeek Harness plugin for gateway routes (OpenCode Go and Command Code): declares the route wire protocol and endpoint (so the API address never has to be typed), detects/reads each route's model catalog and puts the DeepSeek V4.1 models at the front of the OpenCode Go list while completing the Co

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

npx -y @deepseek-ai/dsh plugin --profile web add github:itchenshi/dsh-gateway-models#e3bb5f29968163783607d90846e4cd976900c67f
READMECompatibilityVersions

Description

DeepSeek Harness plugin for gateway routes (OpenCode Go and Command Code): declares the route wire protocol and endpoint (so the API address never has to be typed), detects/reads each route's model catalog and puts the DeepSeek V4.1 models at the front of the OpenCode Go list while completing the Command Code list from its public catalog, supplies the route baseURL and the DeepSeek reasoning compat a catalog-unknown model needs, and attaches a stable per-conversation x-opencode-session header (fixes 400 MissingSessionID) without leaking the internal session id.

Compatibility and provenance

Gateway Models is published as dsh-gateway-models and currently resolves to version 0.2.6. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
10/5/2026

Versions

0.2.6stable
10/5/2026
0.2.5stable
10/5/2026
0.2.3stable
10/5/2026
Show 2 more versionsCollapse versions
0.2.2stable
10/5/2026
0.2.0stable
9/25/2026

Related plugins

Loading related plugins…

Latest
0.2.6
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
any
License
MIT
Source
github
GitHub
★ 0
Weekly downloads
0
Last push
10/5/2026
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.

Experimental Cot Translation Bundle@deepseek-ai/dsh-experimental-cot-translation-bundleExperimental machine translation of expanded reasoningWhale Widgetdsh-whale-widgetDeepSeek balance whale widget in the bottom-right corner of the DSH Web interface: balance/today’s usage/peak-off-peak pricing, customizable bubble click sequence (text/balance/today/peak-off-peak/image/random phrases and parallel weighted selection), per-line styles and fonts, floating quick editinOur Free Model@ebony-vinyl/dsh-our-free-modelOur Free Model — Install this plugin in dsh to use cutting-edge models including DeepSeek V4.1 Flash and Kimi K3, with no login, registration, or API Key required. The desktop version also offers an EAC co-payment channel; starting with 2.0, it supports ten account channels: LobsterAI, Qoder, TRAE, Usage@linxin666/dsh-usageUsage statistics plugin for the dsh web GUI: per-provider balance and coding-plan quota detection plus a live token usage ledger, with a compact price glance card above the sidebar's Settings row

README

dsh-gateway-models

只有宿主半边的插件:给网关路由(OpenCode Go + Command Code)补齐引擎拿不到的协议、地址与模型清单,并给发往 OpenCode 的请求附加会话头。

它做什么

  • 给 opencode-go 路由声明 wire 协议(api: openai-completions),目录外的模型不再报 needs an api。
  • 启动时读引擎自己的模型目录,把 DeepSeek V4.1 模型补进 opencode-go 清单并排到第一位(你自己配过的条目原样保留、顺序不变)。
  • 路由没有 baseURL 时补上 https://opencode.ai/zen/go/v1;你已经配过的地址一个字节都不动。
  • 给 commandcode 路由声明 api 与 baseURL(你只填密钥,不用填地址),并从它公开的目录接口补齐模型清单(只增不改、幂等)。
  • 给发往 OpenCode / OpenCode Go 路由的请求附加 x-opencode-session:默认每个会话一个随机 UUID,绝不把内部会话 ID 发给第三方(mode: 'session-id' 是显式可选)。

安装

  • 用 DSH Ready GUI(推荐):插件随 GUI 内置,打开 GUI → 设置窗口 →「第三方插件」勾选网关路由,装完按提示重启引擎(它没有页面半边)。

  • 其它 DSH 宿主(dsh web / CLI)—— 任选一条:

    # 推荐:直接从 GitHub 装
    dsh plugin --profile web add github:itchenshi/dsh-gateway-models
    
    # 备选:从本仓库 Release 的 tarball 装
    dsh plugin --profile web add https://github.com/itchenshi/dsh-gateway-models/releases/download/v0.2.6/dsh-gateway-models-0.2.6.tgz
    

    两条命令装到的都是这个仓库的完整内容(含 cordis.patch.yml),装完不需要额外配置。

本包还不在 npm 上(registry.npmjs.org 查无此包),只能用上面两种方式装。

使用

没有页面半边,也不需要打开任何界面 —— 重启引擎后它在装配层改配置。你可以在模型选择器里确认:第一项是 DeepSeek V4.1 Flash,llm-pi-ai.providers.opencode-go 里它排在第一位。

Command Code 那一半只对你自己配置过的路由(用户层里有它,比如填过密钥)动手,并按接口地址(api.commandcode.ai)识别路由 —— 路由名叫 commandcode、commandcode-goat 还是别的都行。

配置

cordis.patch.yml 里 id: opencode-go 这一行的 config,全部可选:

键默认说明
providers[opencode, opencode-go]附加会话头的路由名;自定义路由名时加进来
commandcodeProviders无额外的 Command Code 路由 id;仅当某条路由的地址在配置里看不到时才要填
modeuuiduuid = 每会话随机 UUID(安全默认);session-id = 发送内部会话 ID
debugfalsetrue 时记录每个收到会话头的流式调用
debugFile无追加 JSON 日志的路径;仅 $DSH_HOME/logs 或系统临时目录下的路径生效,超过 8 MiB 停止追加

兼容性

  • 需要 DSH >= 0.2.0-rc.2(引擎 0.2.0 更换了设置服务 API,0.1.x 上无法写入路由配置)。
  • Node >= 20。
  • 依赖的是引擎装配层、settings 服务与 llm 事件的公开契约;契约变了引擎会在启动时显式报插件加载失败,而不是静默失效。

常见问题

V4.1 模型没出现? 看宿主启动日志有没有 [gateway-models] active for providers [opencode, opencode-go] with mode uuid (debug=off)。如果打的是 llm.discoverModels() is unavailable 或 could not detect the "opencode-go" model catalog,说明引擎目录读不到 —— 这时插件跳过自动补模型,而不是猜。

Command Code 的模型列表没补齐? 正常会打 [gateway-models] "commandcode": N model(s) from the provider catalog (list X -> Y);如果打的是 could not read the Command Code model catalog (...),说明那次公开目录请求失败了,插件保持原列表不动。

它会读我的 API key 吗? 不会。插件只写路由配置,密钥始终由引擎按 apiKeyEnv 自己解析。唯一出站请求是 GET https://api.commandcode.ai/provider/v1/models(公开、不需要 key、请求里不发 Authorization)。它也不注册任何本机 HTTP 路由。

写配置失败会留下半份吗? 不会。写入走 settings 服务,会重跑引擎的严格校验,校验不过整笔拒绝(日志给出原因)。

开发

node --check lib/index.js
npm test        # node tests/test.mjs:本地行为测试(不联网、不需要引擎)

新增 V4.1 型号只需往 V4_1_MODELS 里加一条;planRouteUpdate() / withV41ModelsFirst() / withV41Defaults() / isV41() 分别负责「写什么」「怎么置顶」「怎么补字段」「是不是 v4.1」,都有单测。

许可

MIT