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.

Models & Usage DSH plugins — Page 44

CATEGORY

Models & Usage DSH plugins

Verified manifests and exact versions in this category.

All categoriesAgents & Orchestration1030Memory & Context1285Coding & Developer Tools2778UI & Customization3631Integrations & Communication1886Vision & Media611Search & Research855Security & Access1258Models & Usage1662Productivity & Workflow2224

All plugins

1665 plugins
Mv1.0.0
@local/model-thinking-levels

Model Thinking Levels

Expose per-model thinking levels (reasoningEfforts) inside the Models settings card for a custom model API route.

models-usage★ 0↓ 0
Updated 9/27/2026
Av0.1.0
dsh-adaptive-max-tokens

Adaptive Max Tokens

Adaptive output-token budgeting for DeepSeek Harness: shrink maxTokens as context fills, restore it after compaction, and retry recognized output-budget overflows once.

models-usage★ 0↓ 0
Updated 9/26/2026
Gv0.1.4
dsh-glm-web

Glm Web

用智谱清言（chatglm.cn）网页版登录态做 DSH 的模型 provider：消耗网页版积分，不消耗开放平台 API 额度。Unofficial Zhipu Qingyan web provider for DeepSeek Harness.

models-usage★ 0↓ 0
Updated 9/26/2026
Pv0.1.0
dsh-plugin-model-hub

Plugin Model Hub

A dsh web plugin: a model/provider configuration page that works over insecure HTTPS (non-loopback browsers) — providers, models, context windows, API keys, per-provider API-error retry/backoff tuning for dsh-llm-retry, and the default Agent model.

models-usagesecurity-access★ 0↓ 0
Updated 9/26/2026
Vv0.3.0
dsh-vision-3090-fix

Vision 3090 Fix

Local reverse proxy plugin for DeepSeek Harness that caps images per request to 1 and drops pinned max_tokens, for self-hosted vLLM backends that reject prompts with more than one image or with prompt_tokens + max_tokens beyond max_model_len

models-usagevision-media★ 0↓ 0
Updated 9/26/2026
Ov0.1.0
dsh-ollama-quota

Ollama Quota

Displays Ollama cloud quota progress in the dsh sidebar: percentage progress bars for session and weekly usage (sidebar.footer.action slot, above Settings), automatically refreshed when a session ends (including heartbeat/webhook sessions).

models-usageui-customization★ 0↓ 0
Updated 9/26/2026
v0.2.8
dsh-subagent-conductor

Subagent Conductor

EN: Settings and default-route layer for official DeepSeek Harness subagents: session/role/global model routing with a composer selector. ZH: 面向 DeepSeek Harness Web 官方子代理的设置与默认路由层：会话/角色/全局模型路由与输入框选择器。

models-usageagents-orchestration★ 0↓ 0
Updated 9/26/2026
Av0.1.0
dsh-anthropic-image-clamp

Anthropic Image Clamp

DeepSeek Harness host plugin: downscale inline images on Anthropic Messages requests so many-image requests never exceed the 2000px per-side limit.

models-usage★ 0↓ 0
Updated 9/26/2026
v0.3.5
dsh-model-health-probe

Model Health Probe

DSH Model Health Check: Adds a “Model Health Check” tab to the session view. Select a target using four rows of buttons: “Provider (grouped by baseURL) → API Type → Model → Route.” Manually send one real raw HTTP test request (supports the three protocols openai-completions / openai-responses / anth

models-usage★ 0↓ 0
Updated 9/25/2026
Kv0.2.3
dsh-kimi-quota

Kimi Quota

Show Kimi Code quota in DeepSeek Harness: a settings section plus a compact composer badge (5h window / weekly)

models-usageui-customization★ 0↓ 0
Updated 9/25/2026
Sv0.1.0
dsh-sub-providers

Sub Providers

Clean-room DSH plugin: use Claude and ChatGPT (Codex) subscriptions as LLM provider routes with OAuth login, streaming, and subscription usage display.

models-usagesecurity-access★ 0↓ 0
Updated 9/25/2026
Uv0.1.2
dsh-usage-plus

Usage Plus

Independent DSH usage and plan-quota plugin with official provider windows, CPAMC / Volcano sources, and a composer-dock quota strip.

models-usageui-customization★ 0↓ 0
Updated 9/25/2026
Tv0.2.1
dsh-team-panel

Team Panel

A DSH plugin for tracking Token usage by project across a team: each person selects a project to monitor in their own DSH, and the plugin records only that project's actual usage and aggregates it on a self-hosted team server. The panel remains at the bottom of the sidebar, showing each person's sel

models-usageui-customization★ 0↓ 0
Updated 9/25/2026
Mv1.2.0
member-model

Member Model

One-shot model route for the Team lead's next fresh teammate. No personal default model.

models-usage★ 0↓ 0
Updated 9/25/2026
Cv0.1.1
dsh-client-ui-route-badge

Client Ui Route Badge

Shows which backend model actually served each request when using the auto route.

models-usageui-customization★ 0↓ 0
Updated 9/25/2026
Mv0.1.3
dsh-model-tester

Model Tester

Model availability tester panel in the Models settings page: one-click real streaming test per added model with TPS / first-token / elapsed metrics

models-usage★ 0↓ 0
Updated 9/25/2026
Jv0.1.0
dsh-jev-router

Jev Router

Jev decision model routing plugin: by default, only adjusts thinking intensity (cache-safe); optional automatic model routing (disabled by default + hard gate on cache cost) · DeepSeek Harness plugin

models-usage★ 0↓ 0
Updated 9/25/2026
Cv1.1.1
dsh-cooldown-retry

Cooldown Retry

Patient auto-retry for DeepSeek Harness: honor the upstream gateway's retry_after_seconds on a 429 capacity cooldown instead of giving up after two fast retries, wait out the mislabeled modality 400 a saturated gateway replays from its client-error circuit, and journal every wait as the built-in's o

models-usage★ 0↓ 0
Updated 9/25/2026
v0.3.0
dsh-zhipu-toolkit

Zhipu Toolkit

Zhipu BigModel/GLM provider toolkit for DeepSeek Harness: dual-endpoint model catalog (Coding Plan + ordinary API), verified thinking-level mapping, live model discovery

models-usage★ 0↓ 0
Updated 9/25/2026
Pv0.3.2
dsh-plugin-model-filter

Plugin Model Filter

Searchable model picker for the DSH chat composer — adds a search/filter box to the model selection menu.

models-usage★ 0↓ 0
Updated 9/24/2026
Pv0.6.7
dsh-provider-switch

Provider Switch

DSH provider management plugin: Adds enable/disable toggles for each provider on the Settings → Models page. When disabled, the provider’s models are hidden from the input model selector and the /model command popover, and conversations will no longer call that provider’s models. Adds keyword search

models-usage★ 0↓ 0
Updated 9/24/2026
Cv0.1.0
dsh-context-length

Context Length

DSH (DeepSeek Harness) plugin: Adds a “Context Length” page under “Settings - Models” for customizing the context window length of a specified model from a specified channel (provider). Changes take effect only after clicking Save.

models-usage★ 0↓ 0
Updated 9/24/2026
v0.1.6
dsh-plugin-model-proxy

Plugin Model Proxy

DSH plugin: per-model proxy routing (http/https/socks5) with a settings UI — e.g. opencode/muse-spark-1.2-contributor needs a proxy while sibling models stay direct

models-usage★ 0↓ 0
Updated 9/24/2026
Sv1.0.1
@datawork-technology/dsh-system-prompt

System Prompt

Per-model system prompt injector for DeepSeek Harness: append custom prompt text to the system prompt for locally configured models, matched by exact provider/model route or a provider-wide provider/* rule, managed from a Settings page and persisted across restarts.

models-usage★ 0↓ 0
Updated 9/24/2026
Sv0.2.0
dsh-session-rescue

Session Rescue

Prevents the DeepSeek Harness cold-start turn stall: after a restart or session load the first turn can block forever while the token meter synchronously replays the whole session log; this plugin compacts the session before its size reaches that danger zone, so the replay stays cheap.

memory-contextmodels-usage★ 0↓ 0
Updated 9/24/2026
Lv0.13.7
dsh-llamacpp-connect

Llamacpp Connect

把本地 llama.cpp 管理器里的模型接入 DeepSeek Harness — bring local llama.cpp manager models into DeepSeek Harness.

models-usage★ 0↓ 0
Updated 9/24/2026
Pv1.0.0
dsh-plugin-usage-stats

Plugin Usage Stats

Whole-corpus token usage statistics: daily and cumulative tokens, cache hit rate, per-model breakdown, and a year-long activity heatmap.

models-usage★ 0↓ 0
Updated 9/24/2026
Dv0.4.1
dsh-dsbal

Dsbal

DeepSeek API balance widget for the DSH web sidebar — CNY balance above Settings, 30s auto refresh, click to refresh, hover details card, threshold warning borders, follows UI language. · DSH 侧边栏 DeepSeek 余额插件：设置按钮上方显示人民币余额，30 秒自动刷新，点击手动刷新，hover 查看余额明细，余额不足阈值告警，文本跟随界面语言。

models-usageui-customization★ 0↓ 0
Updated 9/24/2026
v0.3.13-patched.6
deepseek-harness-wallet-patched

Deepseek Harness Wallet Patched

Local-first DeepSeek Harness control center for balance monitoring, usage accounting, time-aware custom third-party API pricing, a 365-day local token ledger, Z.ai Coding Plan quotas, official pricing sync, encrypted multi-account switching, reminders, recharge, and host-gated session controls.

models-usagesecurity-access★ 0↓ 0
Updated 9/23/2026
v0.1.3
dsh-client-ui-model-clock

Client Ui Model Clock

DeepSeek Harness model deal clock — displays the best-value model in real time, all-day cost heat bands, limited-time deal countdowns, and “Wait a little to save more” scheduling suggestions, while aggregating special offers on resource bundles from various providers. Includes time-window deal data

models-usageui-customization★ 0↓ 0
Updated 9/23/2026
← Previous1…4243444546…56Next →
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in