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.

Subagent Steer — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
S

@momojie-s/dsh-subagent-steer

Subagent Steer

DSH plugin: steer_subagent — insert new instructions into the current turn of a working background subagent (gentle interjection at step boundaries, with optional hard interrupt and restart); the default queuing behavior of send_message remains unchanged

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

npx -y @deepseek-ai/dsh plugin --profile web add github:Momojie-S/dsh-subagent-steer#6fc164fdc34736746abd6758506ed20cdff20285
READMECompatibilityVersions

Compatibility and provenance

Subagent Steer is published as @momojie-s/dsh-subagent-steer and currently resolves to version 0.1.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
9/4/2026

Versions

0.1.0stable
9/4/2026

Related plugins

Loading related plugins…

Latest
0.1.0
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
9/4/2026
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 agents-orchestration.

Headless@deepseek-ai/dsh-headlessThe dsh one-shot bundle: a direct core Agent/Session runner over dsh-base with no Host, HTTP, or browser layerExperimental Agent Team Web Profile@deepseek-ai/dsh-experimental-agent-team-web-profileExperimental Web profile layer for Agent Teams Remote and UI pluginsSubagent Codex@deepseek-ai/dsh-subagent-codexOne-shot Codex subagent provider over the official app-server protocolSubagent Claude Code@deepseek-ai/dsh-subagent-claude-codeOne-shot Claude Code subagent provider over the official Agent SDK

README

@momojie-s/dsh-subagent-steer

给正在工作的后台子agent当轮插入新指示:主 agent 派出去的后台子 agent 干到一半时,把纠偏/补充指令直接送进它当前这一轮(下一个动作步骤前生效),而不是排在整条指令后面干等。官方 send_message 的默认排队行为完全不动,本插件只新增一个 steer_subagent 工具。

环境要求

  • DSH >= 0.1.0-rc.7(已验证至 0.1.1-rc.2;依赖 ctx.subagents.followup / interrupt、ctx.agents 活体注册表、Agent.steer 的 next-step 插入语义)
  • 仅 host 半部,无浏览器资源;Windows/Linux 均可

用法

模型可见一个 steer_subagent 工具(工具名可配)。三种投递行为:

调用方式子 agent 的反应
send_message(官方,不变)排队:把整条当前指令干完,才读新消息
steer_subagent(默认)插话:手头这一步做完(下一次思考前)就读新指示,已完成的步骤保留,剩余步骤按新指示走
steer_subagent + mode: "turn"打断:立即停手,当前轮做到一半的作废,新指示作为全新一轮开始(已排队未消费的消息保留)

参数:

参数必填说明
subagent_id✅后台子 agent 启动时返回的 id
message✅要插入的指示
mode—'step'(默认,温和插话)/ 'turn'(硬打断重开);省略取默认

行为要点:

  • 目标不在线时自动回退:子 agent 空闲或只存于存储时,自动走与 send_message 相同的唤醒/冷恢复路径,FIFO 顺序不变(工具结果会注明实际走的路线 route)
  • 授权:只允许父 agent 插话自己的直接后台子 agent;更深层后代仍是 interrupt_agent 的候选
  • 插话即时生效,回信照旧:插话立刻进入子 agent 当前轮;但它的回复与结算通知仍按 DSH 现有机制跨轮送达
  • 适用对象:可继续(continuable)后台子 agent;一次性子 agent 若在读到插话前就跑完,插话会随收件箱清理被丢弃

安装

dsh plugin --profile web add github:Momojie-S/dsh-subagent-steer
# 首次按 pnpm 提示在 profile 的 pnpm-workspace.yaml 加 allowBuilds 授权构建

本仓开发机:dsh plugin --profile web add <本目录路径>,然后重启 DSH。

配置

patch config 字段(cordis.patch.yml 行内覆盖):

字段类型默认说明
toolNamestringsteer_subagent模型可见工具名(多实例需互异)

验证

主会话派一个慢速后台子 agent(如"从 1 数到 12,每步等 4 秒"),数到 3 左右时调用:

steer_subagent  subagent_id: <id>
  message: 当前步完成后立即停止计数,汇报你完成到第几步,并逐字引述本条指示

子 agent 的最终汇报引述了指示、完成步数停在 3 左右、剩余步骤未执行 = 全链路正常(实测 12 步任务在第 3 步被插话后,第 4~12 步零执行)。

设计取舍与机制详见 docs/design/overview.md。