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.

Community Feature Requests — Plugin Hub News
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in

COMMUNITY CHALLENGES

Find your next challenge.

Real problems from the community. Find one that sparks an idea, explore the discussion, and build something people need.

Showing 61–80 of 130 community requestsUpdated hourly from community discussions
All updatesDSH ReleasesFeature Requests
All3 days7 days14 days
TopNew
OpenShippedSolved by pluginWon't fixAll statuses

Window Sep 6, 2026 – Sep 20, 2026

#Problems worth solvingCommunity demand
61
HUB
hackernews4

Create a unified skills registry and standard

Provide a centralized, standardized registry for creating, discovering, versioning, distributing, and synchronizing reusable agent skills across compatible agents and developers.

Explore challenge ↗

The problem

Provide a centralized, standardized registry for creating, discovering, versioning, distributing, and synchronizing reusable agent skills across compatible agents and developers.

Possibly addressed by:Skills Hub · Skill Manage · Skills Anywhere

Start with these discussions

  • Comment on: Claude Code now reads AGENTS.md if there is no Claude.mdhackernews · ▲0 · 0 comments
  • Comment on: Ask HN: How do you manage skills files?hackernews · ▲0 · 0 comments
  • Comment on: Ask HN: How do you manage skills files?hackernews · ▲0 · 0 comments
Last active Sep 18, 2026
4 authors4 posts · 0 reactions0 comments
62
PLUGIN
2

Configure reasoning for auxiliary calls

Allow compaction and session-title calls to use independent reasoning settings, defaulting to no reasoning when unspecified while preserving ordinary route defaults. Ensure title generation has a deterministic fallback when a reasoning model returns no text.

Explore challenge ↗

The problem

Allow compaction and session-title calls to use independent reasoning settings, defaulting to no reasoning when unspecified while preserving ordinary route defaults. Ensure title generation has a deterministic fallback when a reasoning model returns no text.

Possibly addressed by:Reasoning Settings · Plugin Reasoning Efforts · Reasoning Level

Start with these discussions

  • Session titles silently fail to generate on reasoning (thinking) models — no config way to disable reasoning for the auxiliary title callgithub · ▲2 · 4 comments
  • Honor `purpose` when an auxiliary call names no reasoning levelgithub · ▲1 · 5 comments
Last active Sep 18, 2026
63
HARNESS
1hackernews1

Add audio notifications for agent state changes

Provide built-in or reusable audio and browser notifications for important agent states, including waiting, blocked, and round completion, so users can monitor agents without continuously watching the harness.

Explore challenge ↗

The problem

Provide built-in or reusable audio and browser notifications for important agent states, including waiting, blocked, and round completion, so users can monitor agents without continuously watching the harness.

Possibly addressed by:Chime Alerts · Notify · Agent Notify

Start with these discussions

  • 浏览器声音+notification 提示完成一轮github · ▲1 · 0 comments
  • Comment on: Show HN: Share your AI Setup, Learn from othershackernews · ▲0 · 0 comments
Last active Sep 18, 2026
2 authors
64
HARNESS
1hackernews1

Enable browser-based computer use

Allow agents to browse live web pages, click elements, search, and fill forms, including through the user's existing browser session and logged-in accounts. This would support end-to-end web tasks and business workflows directly from the harness.

Explore challenge ↗

The problem

Allow agents to browse live web pages, click elements, search, and fill forms, including through the user's existing browser session and logged-in accounts. This would support end-to-end web tasks and business workflows directly from the harness.

Possibly addressed by:Browser Use · Browser Ctrl · Browser Agent

Start with these discussions

  • Navegación web interactiva (tipo ChatGPT Agent) en DPSEEKgithub · ▲1 · 0 comments
  • Comment on: DeepSeek launching v4.1 flash cheaper and more capable than v4 prohackernews · ▲0 · 0 comments
Last active Sep 13, 2026
65▼15
PLUGIN
2

Signal and manage breaking package removals

Provide clear deprecation and removal signals for packages and APIs, including metadata, compatibility guards or deprecated stubs, removal records, release notes, and migration guidance. Help consumers detect upcoming removals and avoid misleading runtime failures when incompatible versions are resolved.

Explore challenge ↗

The problem

Provide clear deprecation and removal signals for packages and APIs, including metadata, compatibility guards or deprecated stubs, removal records, release notes, and migration guidance. Help consumers detect upcoming removals and avoid misleading runtime failures when incompatible versions are resolved.

Start with these discussions

  • Published packages have no removal signal — two public surfaces removed in six days, one still installs clean from npmgithub · ▲2 · 1 comments
  • [dsh-settings] settingsNamespace export removed in 0.1.2-alpha.x without migration cue — misleading 'does not provide an export' crashgithub · ▲1 · 1 comments
Last active Sep 6, 2026
1 authors1 posts · 2 reactions1 comments
66▼27
HARNESS
5

Unify the Plugin Architecture

Build a minimal, flexible plugin-based core that supports user-created plugins across GUI and CLI environments, while implementing DSH’s own desktop capabilities as plugins.

Explore challenge ↗

The problem

Build a minimal, flexible plugin-based core that supports user-created plugins across GUI and CLI environments, while implementing DSH’s own desktop capabilities as plugins.

Possibly addressed by:Deepseek Harness Plugins · Plugin Optimization · Plugin Hub

Start with these discussions

  • @nachooolk @emanueledpt Also do you feel like it would be good if there was a plugin/customisation system so that you cax · ▲0 · 2 comments
  • @Teknium @honchodotdev @mem0ai Doesn't deepseek harness architecture serve as a better guide? If you look at it, you onx · ▲0 · 0 comments
  • I was thinking for hermes like deepseek harness concept. Means everything is plugin so hermes imo should able to createx · ▲0 · 0 comments
Last active Sep 19, 2026
3 authors3 posts · 0 reactions0 comments
67
HARNESS
hackernews3

Provide reusable on-demand skills for custom workflows

Let users package custom agent flows as reusable skills and define repeatable, on-demand tasks outside AGENTS.md. This enables workflows such as migrations or component generation without unnecessarily polluting the agent’s context.

Explore challenge ↗

The problem

Let users package custom agent flows as reusable skills and define repeatable, on-demand tasks outside AGENTS.md. This enables workflows such as migrations or component generation without unnecessarily polluting the agent’s context.

Possibly addressed by:Agent Useful Skills · Skill Creator · Agent Skills

Start with these discussions

  • Comment on: Claude Code now reads AGENTS.md if there is no Claude.mdhackernews · ▲0 · 0 comments
  • Comment on: OpenAI Agents APIhackernews · ▲0 · 0 comments
  • Comment on: Ask HN: How do you manage skills files?hackernews · ▲0 · 0 comments
Last active Sep 18, 2026
3 authors3 posts · 0 reactions0 comments
68▲13
PLUGIN
2

Detect and block repeated tool-call loops

Detect repeated, cyclic, or near-identical tool-call patterns across configurable windows and turn counts, then block or interrupt them and redirect the agent toward synthesis, clarification, or a different strategy.

Explore challenge ↗

The problem

Detect repeated, cyclic, or near-identical tool-call patterns across configurable windows and turn counts, then block or interrupt them and redirect the agent toward synthesis, clarification, or a different strategy.

Possibly addressed by:Loop Continue · Tool Failure Circuit Breaker · Repeat Stop

Start with these discussions

  • Repeat-tool reminder detects loops but does not reliably break themgithub · ▲1 · 3 comments
  • [Idea] repeat-tool-reminder 抓不住调用签名周期循环(A→B→C→A)——只统计连续相同签名,周期循环永远不触发github · ▲0 · 0 comments
Last active Sep 12, 2026
1 authors
69▼50
HARNESS
1hackernews81

Implement policy-based model routing

Add configurable routing that selects models by task, capabilities, cost, priority, health, quotas, and fallback policies, including delegation to smaller specialized models when appropriate.

Explore challenge ↗

The problem

Add configurable routing that selects models by task, capabilities, cost, priority, health, quotas, and fallback policies, including delegation to smaller specialized models when appropriate.

Possibly addressed by:Model Router · Model Tier · Auto Model Router

Start with these discussions

  • Design: Policy-driven multi-source quota router for DSH
70
HARNESS
2

Package macOS Apps Locally Without Signing

Provide a command or workflow to package the macOS desktop app locally without a developer account or signing certificates, and without modifying source code.

Explore challenge ↗

The problem

Provide a command or workflow to package the macOS desktop app locally without a developer account or signing certificates, and without modifying source code.

Possibly addressed by:Desktop Packager · Desktop Kit · Deepseek Harness Desktop

Start with these discussions

  • 跪求 mac 桌面版本打包支持本地打包,不要求证书啊github · ▲1 · 0 comments
  • MacOs 打包桌面版本希望支持不要开发者账号的的本地版本命令github · ▲1 · 0 comments
Last active Sep 11, 2026
1 authors2 posts · 2 reactions0 comments
71
HARNESS
2

Add Opt-In Live Delta Streaming

Expose an opt-in, non-durable SDK stream for assistant text, reasoning, and tool-call deltas as they arrive, including through the JSON-RPC protocol. Keep durable assistant messages and session events authoritative for persistence and replay.

Explore challenge ↗

The problem

Expose an opt-in, non-durable SDK stream for assistant text, reasoning, and tool-call deltas as they arrive, including through the JSON-RPC protocol. Keep durable assistant messages and session events authoritative for persistence and replay.

Possibly addressed by:Stream Upload · Event Relay

Start with these discussions

  • [SDK/讨论] JSON-RPC wire 能否增加可选的实时 assistant-stream 通知?Web UI 已有进程内实时通道,out-of-process SDK 消费端只能拿 durable 结算github · ▲1 · 0 comments
  • SDK runtime: request for an opt-in best-effort live streaming channel alongside durable eventsgithub · ▲1 · 0 comments
Last active Sep 11, 2026
1 authors2 posts · 2 reactions0 comments
72▼6
PLUGIN
3

Allow selecting agent presets when starting sessions

Allow session creation to select an agent preset through workspace/session APIs and headless CLI, including loading presets from .dsh/.agent-presets/ with an --agent-preset option.

Explore challenge ↗

The problem

Allow session creation to select an agent preset through workspace/session APIs and headless CLI, including loading presets from .dsh/.agent-presets/ with an --agent-preset option.

Possibly addressed by:Preset · Workspace Presets · Agent Preset Switcher

Start with these discussions

  • 建议:在 uiWorkspace.openWorkspace() / sessions.create() 中支持指定 agent preset(模式)github · ▲1 · 0 comments
  • 建议:在 uiWorkspace.openWorkspace() / sessions.create() 中支持指定 agent preset(模式)github · ▲1 · 0 comments
  • 建议:headless 支持 --agent-preset 旗标(预设能力对齐 Web GUI)github · ▲1 · 0 comments
73
PLUGIN
2

Map pi-ai reasoning tokens

Map the pi-ai provider’s reasoning-token count to `TokenUsage.reasoningTokens` while preserving output-token counts. Omit the field when the count is zero so reasoning and answer usage can be separated consistently across transcripts, context meters, and per-turn accounting.

Explore challenge ↗

The problem

Map the pi-ai provider’s reasoning-token count to `TokenUsage.reasoningTokens` while preserving output-token counts. Omit the field when the count is zero so reasoning and answer usage can be separated consistently across transcripts, context meters, and per-turn accounting.

Possibly addressed by:Plugin Reasoning Efforts · Thinking Token Stat · Custom Provider Reasoning

Start with these discussions

  • llm-pi-ai drops the reasoning-token breakdown pi-ai already reportsgithub · ▲2 · 2 comments
  • Map pi-ai's reasoning usage into `reasoningTokens`github · ▲1 · 2 comments
Last active Sep 18, 2026
74▼1
PLUGIN
2

Support configuring additional sandbox writable roots

Allow sandbox policy configuration to specify additional writable root directories, such as SDK installation paths, so toolchain caches can be written without expanding the workspace or requiring repeated permission escalation.

Explore challenge ↗

The problem

Allow sandbox policy configuration to specify additional writable root directories, such as SDK installation paths, so toolchain caches can be written without expanding the workspace or requiring repeated permission escalation.

Possibly addressed by:Write Protect · Sandbox Allowlist · Plugin Multi Root Workspace

Start with these discussions

  • Sandbox writableRoots hardcoded — no way to add extra writable dirs for SDK toolchains / 沙箱可写根目录无法附加(SDK工具链)github · ▲2 · 1 comments
  • [Feature request] workspace-write 增加 additionalWritableRoots:支持工作区外共享工具链缓存的可写配置github · ▲1 · 2 comments
Last active Sep 13, 2026
75▼31
HARNESS
2

Browse and view workspace files in the web UI

Add a web-based workspace browser that supports directory navigation, source-file viewing, and inspection of code changes. It should work well for remote or headless deployments and make agent edits accessible without a local editor.

Explore challenge ↗

The problem

Add a web-based workspace browser that supports directory navigation, source-file viewing, and inspection of code changes. It should work well for remote or headless deployments and make agent edits accessible without a local editor.

Possibly addressed by:Solution Explorer · Session Diff · Code Workbench

Start with these discussions

  • 可以提供类似opencode的web页面中源代码文件查看功能github · ▲1 · 4 comments
  • DSH的webUI看不到本地文件内容真的挺难受的github · ▲1 · 0 comments
Last active Sep 20, 2026
1 authors1 posts · 1 reactions2 comments
76
HARNESS
3

Implement scoped, pluggable agent memory

Provide a backend-agnostic memory service or plugin with session, agent, and tenant isolation, automatic capture and recall, profile injection, safe lifecycle handling, and support for importing existing memories. Document secure multi-tenant integration and prevent races or duplicate writes during cleanup and migration.

Explore challenge ↗

The problem

Provide a backend-agnostic memory service or plugin with session, agent, and tenant isolation, automatic capture and recall, profile injection, safe lifecycle handling, and support for importing existing memories. Document secure multi-tenant integration and prevent races or duplicate writes during cleanup and migration.

Possibly addressed by:Plugin Memory · Agent Memory · Agentmemory

Start with these discussions

  • dsh-plugin: Native OpenViking memory for DeepSeek Harness (automatic session capture + memory tools + profile injection)github · ▲2 · 5 comments
  • 是否可以支持多租户记忆隔离,提供详细的api文档github · ▲1 · 3 comments
  • 给 #14 补点料:两个现成记忆实现里踩过的坑
77▲17
HARNESS
1hackernews1

Add persistent quota usage tracking and display

Track provider and model quota usage across configured rolling or billing windows and present it through an always-visible graphical dashboard. Include progress indicators and warnings so users can monitor consumption and avoid exhausting their allowance.

Explore challenge ↗

The problem

Track provider and model quota usage across configured rolling or billing windows and present it through an always-visible graphical dashboard. Include progress indicators and warnings so users can monitor consumption and avoid exhausting their allowance.

Possibly addressed by:Quota Monitor · Quota Viewer · Deepseek Harness Quota Monitor

Start with these discussions

  • [Idea] First-class subscription quota windows (5-hour / weekly / monthly) for Coding Plan providersgithub · ▲2 · 0 comments
  • Comment on: Show HN: My Claude quota ran out in 10 minutes, so I made a tool to find out why
78▼53
HARNESS
3

Support configurable inference tiers

Add provider-aware, user-selectable compute or service tiers with distinct cost, latency, and capability trade-offs, optional automatic selection, and end-to-end propagation through model and session configuration. Preserve defaults, persist choices, expose usage metadata, and never silently downgrade unsupported tiers.

Explore challenge ↗

The problem

Add provider-aware, user-selectable compute or service tiers with distinct cost, latency, and capability trade-offs, optional automatic selection, and end-to-end propagation through model and session configuration. Preserve defaults, persist choices, expose usage metadata, and never silently downgrade unsupported tiers.

Possibly addressed by:Model Tier · Tier Router · Better Model Provider

Start with these discussions

  • Feature request: end-to-end support for OpenAI Codex Standard/Fast service tiersgithub · ▲1 · 1 comments
  • User-selectable compute tiers with different token pricing for DeepSeek V4.1 Flashgithub · ▲1 · 0 comments
79▼17
HARNESS
2

Reject Unsupported Node.js Versions Before Startup

Add an early Node.js version guard to the CLI and build entry points, declare the supported engine range, and return a clear error with a nonzero exit code for unsupported versions.

Explore challenge ↗

The problem

Add an early Node.js version guard to the CLI and build entry points, declare the supported engine range, and return a clear error with a nonzero exit code for unsupported versions.

Possibly addressed by:Check For Updates · Install Guard · Version Update

Start with these discussions

  • Fix(cli) Proposal: 让入口守卫在 Node 24.0–24.1 上失败可见,并把它们移出 engines 范围github · ▲1 · 1 comments
  • Node 20 causes npx @deepseek-ai/dsh web to hang without a clear version errorgithub · ▲1 · 0 comments
Last active Sep 14, 2026
1 authors1 posts · 1 reactions1 comments
80▼32
PLUGIN
2

Persist settings for trusted remote browsers

Allow UI and settings-plugin preferences to use host persistence for remote browsers that pass the trusted-host fence, including declared hosts, LAN names, and reverse proxies.

Explore challenge ↗

The problem

Allow UI and settings-plugin preferences to use host persistence for remote browsers that pass the trusted-host fence, including declared hosts, LAN names, and reverse proxies.

Possibly addressed by:Settings Remote · Trusted Host Proxy 403 Fix · Remote Settings Compat

Start with these discussions

  • web settings: Host persistence is loopback-only — remote browsers get "settings are unavailable in this browser"github · ▲1 · 1 comments
  • UI appearance / settings changes made over a non-loopback (trusted remote) origin do not persist — `persistence` is hard-gated to loopback onlygithub · ▲1 · 1 comments
Last active Sep 6, 2026
← PreviousPage 4 of 7Next →
1 authors1 posts · 1 reactions5 comments
2 posts · 1 reactions
0 comments
2 authors2 posts · 1 reactions0 comments
1 posts · 1 reactions
3 comments
github · ▲1 · 0 comments
  • Comment on: >10x More Efficient Pretraininghackernews · ▲0 · 0 comments
  • Comment on: DeepSeek launching v4.1 flash cheaper and more capable than v4 prohackernews · ▲0 · 0 comments
  • Last active Sep 10, 2026
    3 authors3 posts · 0 reactions0 comments
    Last active Sep 10, 2026
    1 authors2 posts · 2 reactions0 comments
    1 authors
    1 posts · 1 reactions
    2 comments
    1 authors1 posts · 1 reactions2 comments
    github · ▲1 · 2 comments
    Last active Sep 8, 2026
    1 authors1 posts · 1 reactions2 comments
    hackernews · ▲0 · 0 comments
    Last active Sep 17, 2026
    1 authors1 posts · 2 reactions0 comments
  • Proposal: configurable OpenAI-compatible service tier for llm-pi-aigithub · ▲1 · 0 comments
  • Last active Sep 14, 2026
    1 authors1 posts · 1 reactions1 comments
    1 authors1 posts · 1 reactions1 comments