DeepSeek Harness Plugin Hub

发布与管理完整 Harness Profiles,发现适合你的插件。

探索

插件目录环境预设文档中心动态

社区

发布插件联系我们报告问题

相关链接

Plugin Hub GitHubDeepSeek Harness 官方项目系统状态隐私说明
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

独立、非官方社区项目,与 DeepSeek 官方无隶属、授权或背书关系。

Ultracode — DeepSeek Harness 插件(DSH Plugin)
DeepSeek Harness Plugin Hub
ProfilesPlugins分类动态文档登录管理 Profiles
ProfilesPlugins分类动态文档登录
← Plugins
U

dsh-ultracode

Ultracode

DeepSeek Harness 的 Ultra 模式:按会话设置的工作强度等级,将推理固定为最高级别并激活编排策略部分,同时在 Web UI 中显示 composer ULTRA 标签和彩虹边框

插件会安装到这里;不确定时保持 web。

npx -y @deepseek-ai/dsh plugin --profile web add dsh-ultracode@0.1.0
README兼容性版本

兼容性与来源证明

Ultracode 以 dsh-ultracode 发布,当前版本为 0.1.0。Plugin Hub 会校验它的 manifest,并保存精确安装来源,便于复现安装结果。

DSH 兼容范围
*
运行环境
web
发布来源
npm
Registry 更新时间
2026/8/20

版本

0.1.0stable
2026/8/20

相关插件

正在加载相关插件…

最新版
0.1.0
DSH
*
HMR
重启进程
Tree shaking
未声明可安全裁剪
解包体积
1 MB
文件数
18
Surface
web
许可证
MIT
发布源
npm
GitHub
★ 0
周下载
0
最近提交
2026/9/14
查看源码 ↗
README Badge

点击下方 Badge 复制 Markdown,粘贴到 README 即可。

这是你的 Plugin?认领权益 · 优先安全扫描

验证 package.json 声明的 GitHub 仓库,即可管理这个公开页面。认领后,Hub 会优先安排当前版本的安全扫描,并在通过后公开展示结果。

认领这个 Plugin →
报告问题

相关插件

继续浏览 agents-orchestration 分类下经过校验的插件。

Headless@deepseek-ai/dsh-headlessdsh one-shot bundle:基于 dsh-base 的直接核心 Agent/Session 运行器,不包含 Host、HTTP 或浏览器层Experimental Agent Team Web Profile@deepseek-ai/dsh-experimental-agent-team-web-profile用于 Agent Teams Remote 和 UI 插件的实验性 Web 配置层Subagent Codex@deepseek-ai/dsh-subagent-codex基于官方 app-server 协议的一次性 Codex 子代理提供程序Subagent Claude Code@deepseek-ai/dsh-subagent-claude-code基于官方 Agent SDK 的一次性 Claude Code 子代理提供方

README

English · 简体中文

DeepSeek Harness Ultracode

One switch. Maximum depth.

dsh-ultracode adds an ultra tier to DeepSeek Harness: one per-session switch that pins every model request to the serving model's deepest reasoning effort and activates a standing orchestration policy — workflow fan-out by default on substantive tasks, adversarial verification before reporting, context hygiene.

Modeled on Claude Code's ultracode tier: the tier's substance is the standing orchestration policy, not a deeper reasoning parameter. Works on any provider whose adapter declares efforts from the shared vocabulary — DeepSeek (max), GLM / Kimi routes (high or max, whichever the model declares), or fully custom OpenAI-compatible routes.

Toggling ultra mode from the composer: the ULTRA chip switches the tier, the effort picker follows, and the composer card carries an animated rainbow border

Toggle it from the composer: a ULTRA chip appears next to the model selector, and while active the composer card carries an animated rainbow border.

What it changes

CapabilityWhat it changes
Per-session effort tierOne switch pins every model request to the serving model's deepest declared effort (auto), or a literal configured value.
Orchestration policyA deployment-owned prompt section rides the system prompt while active: standing workflow opt-in, adversarial verification, narrow-the-wave retries on provider limits.
Composer chip + rainbow borderA ULTRA chip in the composer bar; the composer card glows while the tier is in force (static under prefers-reduced-motion).
Picker consistencyToggling re-syncs the native effort picker through the public models API, so the display always matches the pinned reality.
Delegation-awareSpawn subagents inherit the effective state through the delegation chain; forks inherit through their seeded prefix and freeze there.

Install

[!NOTE] Requires an existing DeepSeek Harness installation (@deepseek-ai/dsh-* >= 0.1.0-rc.6).

npm

dsh plugin --profile <name> add dsh-ultracode

Build from source

git clone https://github.com/BackMountainBird/dsh-ultracode.git
cd dsh-ultracode
pnpm install
pnpm build
dsh plugin --profile <name> add .

Run pnpm build again after changing the source. The local plugin install remains linked to this checkout.

Validate the composed profile and restart DSH:

dsh --profile <name> --dump-config
dsh --profile <name> web

Then switch the tier on — from the composer chip, or directly:

/ultra

Use ultra mode to audit every integration point of this repository against its pinned dependencies, with parallel verification.

Leave with /ultra off (or click the chip again).

How it works

  1. The switch is one command. /ultra (and /ultra off) execute through the harness command runtime, which appends the command/run session event before the handler runs. That event is the durable state — the plugin defines no session event of its own, so resume and replay restore it for free, and the last command wins.
  2. The policy rides the system prompt. While active, an ultra:policy section renders into every request. The text is deployment-owned config; the default is the orchestration policy (standing workflow opt-in, adversarial verification, context hygiene, narrow the wave when a provider's concurrency limit times parallel subagents out).
  3. The effort is pinned per request. An agent/request waterfall listener — prepended, so it stays outside the per-agent model-selection listener that would strip it — replaces each request's reasoning effort. auto (default) resolves the deepest effort the serving model declares under the shared off…max vocabulary; a literal value pins exactly and fails loud if the adapter does not declare it.
  4. Delegation inherits it. Spawn children walk the delegation chain (each child's parentSession, resolved against the live session store) and get the pin and policy too — the effort resolving against the child's model. Fork children inherit only through their seeded completed-turn prefix; a parent's later switch never reaches an already-spawned branch.
  5. The model is told. Each genuine switch injects one plugin-source notice user message, so the model learns the change without diffing prompt sections.
  6. The picker follows. The chip's toggle also re-submits the session's model selection through the public models API (deepest effort on switch, provider default on exit), so the native effort picker shows the tier actually in force. The request-side pin remains the guarantee.
  7. The web half is one client bundle. A ULTRA chip occupies the composer's conversation.input.right seat; state rides the host-computed ultra projection (zero client-side ultra state), and while active the chip marks the composer card — an injected stylesheet renders the rainbow border. The bundle ships as a closure factory (window.__ModuleLoader__.load) with platform externals resolved through the loader's module table.

Configuration

Profile row dsh-ultracode:

- id: dsh-ultracode
  config:
    effort: auto
FieldDefaultMeaning
sectionbuilt-in policy textRendered as the ultra:policy prompt section while active
effortautoauto pins each request to the deepest effort the serving model declares; a literal value must be one of the serving adapter's declared efforts (fails loud otherwise)
promptSectionOrder120Prompt-section order

Boundaries

  • State is the harness-owned command/run event: no custom session events, and the tool catalog never changes across modes (request-cache stability).
  • Inheritance only ever adds: subagent sessions have no user command surface, so children cannot opt out.
  • The request-side pin is the truth; the picker sync is display consistency — a manual pick made mid-ultra realigns on the next toggle, while requests keep running pinned.
  • Fan-out capacity belongs to the provider: the default policy tells the model to retry timed-out parallel work as a narrower wave instead of abandoning it.
  • Ultra does not touch plan mode: the two are independent logged states and compose freely (/plan + /ultra = max-effort planning under plan's read-only constraints).

Development

pnpm install
pnpm run typecheck
pnpm run build
pnpm run verify

Host half in src/ (command, effort pin, prompt section, session projection, delegation-chain walk); browser half in src/client/ (composer chip, rainbow stylesheet, picker sync). Peer floors are @deepseek-ai/dsh-*@^0.1.0-rc.6; devDependencies pin the floor so development cannot drift onto newer-only APIs.

License

MIT