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。

Aigc Radar — DeepSeek Harness 插件(DSH Plugin)
← Plugins

dsh-aigc-radar

Aigc Radar

AIGC Radar for DeepSeek Harness — 精选 AI 项目搜索,提供原生搜索卡片,由 AIGC_NEWS 库提供支持

插件会安装到这里;不确定时保持 web。

npx -y @deepseek-ai/dsh plugin --profile web add github:Kaixxrua/dsh-aigc-radar#d2ece97b883d005c8e8b587bb93699c111b1d435
README兼容性版本

兼容性与来源证明

Aigc Radar 以 dsh-aigc-radar 发布,当前版本为 0.2.5。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
any
发布来源
github
Registry 更新时间
2026/9/20

版本

0.2.5stable
2026/9/3
0.2.4stable
2026/9/1
0.2.3stable
2026/8/22
查看其余 2 个版本收起版本
0.2.2stable
2026/8/21
0.2.1stable
2026/8/20

相关插件

正在加载相关插件…

最新版
0.2.5
DSH
*
HMR
重启进程
Tree shaking
未声明可安全裁剪
解包体积
未提供
文件数
未提供
Surface
any
许可证
MIT
发布源
github
GitHub
★ 4
周下载
124
最近提交
2026/9/3
查看源码 ↗项目主页 ↗
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
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

相关插件

继续浏览 search-research 分类下经过校验的插件。

Browser Skill Dsh Plugin@wxg-prc-cpg/browser-skill-dsh-plugin向模型提供 BrowserSkill 浏览器自动化(browser_* 工具)的 DeepSeek Harness 工具插件Weknora@wxg-prc-cpg/dsh-weknora适用于 DeepSeek Harness (dsh) 的 WeKnora 知识检索工具:通过自有知识库进行语义搜索、文档阅读以及 RAG/代理回答。Free Searchdsh-free-searchDeepSeek Harness 的免费网页搜索:13 个引擎(Bing/DuckDuckGo/AnySearch/SearXNG/Exa/Tavily/Keenable/Firecrawl 无需密钥;Parallel/Perplexity/SerpBase/DeepSeek 需要密钥),支持时间筛选、平台搜索和 web_fetch,并提供网页设置界面。Find Plugindsh-find-plugin在代理中查找 DeepSeek Harness 插件——实时搜索 GitHub 上的 dsh-plugin 主题,并按星标数排序。

README

dsh-aigc-radar

English | 简体中文

AIGC Radar project search for DeepSeek Harness (dsh).

Stop rebuilding what already exists. While you plan and implement, the agent proactively checks the curated AIGC Radar library for mature, battle-tested projects that already solve your problem — before you write a line of code. Results render as native search cards in the dsh Web UI — not raw markdown — and survive session replay.

search_ai_projects rendering as a native search card in the dsh Web UI

What you get

ToolWhat it does
search_ai_projectsSearches the curated AIGC Radar library: GitHub projects above a 500-star floor, enriched with categories, bilingual (zh/en) tags and descriptions, and daily star-growth metrics
get_project_categoriesLists the category taxonomy (categories + subcategory counts) for filter discovery

Two routing layers make discovery automatic rather than opt-in:

  • Explicit discovery — ask "找个能做 deep research 的开源框架" and the agent comes back with starred, categorized results, no "use the tool" needed
  • Proactive reuse check — before the agent implements a major module or subsystem (auth, payments, workflow engines, search/indexing, protocol implementations, end-to-end RAG/Agent pipelines…), or recommends the architecture / technology / build-vs-reuse approach for one, it runs one library check on its own initiative, so mature alternatives surface before anyone rebuilds them. Narrow work (bug fixes, renames, styling, CRUD) is deliberately excluded

Why a native plugin instead of the MCP server?

AIGC Radar also ships as an MCP server — and this plugin now rides that same MCP endpoint (POST /api/mcp), so every call counts against the same rate limits and quotas. What the plugin adds on top:

  • Native web search cards — structured sources render as cards in the Web UI and are rebuilt faithfully on session replay (presentationMeta), which the MCP transport cannot express
  • Typed canonical output — the result is one validated JSON value, so Code Mode can compose it programmatically (await tools.search_ai_projects({ q: 'mcp' })) with full type inference
  • First-party prompt routing — the discovery-routing guidance lives in system-prompt assembly, not in MCP instructions that clients may truncate

Measured performance

The search tool is a single HTTPS call to the AIGC Radar public edge — the numbers below measure that full path, taken 2026-08-18 from a China home-broadband connection (GeoDNS → CN edge) with scripts/benchmark-search.sh (10 representative zh/en queries × 3 trials against https://aigcnews.cn/api/mcp):

MetricValue
Search latency p50355 ms
Search latency p95810 ms
Curated projects served18,426 — every one above the 500-star admission bar
Taxonomy11 top-level categories, bilingual zh/en tags and descriptions

Index-quality benchmarks measured on the MCP interface — 3.5× faster than WebSearch end-to-end (7.4 s vs 25.8 s median) and 98.3% first-tool routing accuracy (59/60 trials) — live in the main repo's benchmark section; this plugin serves the same dataset over the same API, so those numbers carry over.

Install

Requires dsh (npx @deepseek-ai/dsh web).

Recommended — install the prebuilt package from npm:

dsh plugin --profile web add dsh-aigc-radar

For a reproducible install, pin the published release:

dsh plugin --profile web add dsh-aigc-radar@0.2.2

Source fallback — install from GitHub:

dsh plugin --profile web add github:Kaixxrua/dsh-aigc-radar

Git installs run the package's prepare build, which pnpm ≥10 refuses until you allow it: copy the package key pnpm prints into the profile's pnpm-workspace.yaml:

allowBuilds:
  dsh-aigc-radar: true

Then re-run the GitHub add. Pin a commit (github:Kaixxrua/dsh-aigc-radar#<sha>) if you want the source install to be immutable.

Verify without booting:

dsh --profile web --dump-config   # shows a "# == dsh-aigc-radar" layer

Updates

Update an npm install within its declared semver range:

dsh plugin --profile web update dsh-aigc-radar

Restart dsh afterwards to load the new version — a running dsh process is not hot-swapped.

Since 0.2.2, the plugin itself also notices when a newer release exists: on dsh start it makes one read-only npm registry check per process, and the agent relays the exact update command at the start of the next turn. The plugin never modifies its own installation. Set updateCheck: false in the plugin config to disable the check.

If you pinned an exact version (for example dsh-aigc-radar@0.2.1), name the target version explicitly:

dsh plugin --profile web add dsh-aigc-radar@<version>

Git/Git SHA or branch, file/link, workspace, tarball, and local-path installs stay manual and are never changed by these commands. Startup-time automatic update checks are not yet available in released dsh builds; until then, the commands above are the update path.

Recommended: register on the origin site and grab a free token. The plugin works anonymously out of the box, but anonymous calls share a 30-calls/day per-IP bucket. A free account at aigcnews.cn gets you 60 calls/day plus a shared 1,200-call rolling-30-day quota; a member account gets 400/day plus 8,000 per rolling 30 days. Create an MCP token on the /mcp page (no special scopes needed for search) and paste it as mcpToken in the config below. See Quotas and the MCP token.

Configure

Defaults point at the public deployment. Override the row from your profile's cordis.patch.yml (a patch replaces the row's whole config):

- replace:
    - id: aigc-radar
      name: dsh-aigc-radar
      config:
        apiBase: 'https://aigcnews.cn'   # or your self-hosted AIGC_NEWS origin
        mcpToken: ''                     # MCP token from {apiBase}/mcp; empty = anonymous
        timeoutMs: 20000
        maxPageSize: 10                  # capped at 20 by the MCP contract
        updateCheck: true                # set false to skip the once-per-process release check

Quotas and the MCP token

Every call lands in the MCP endpoint's quota domain — anonymous callers are bucketed per IP, token callers per account:

CallerQuotaWindow
Anonymous (no mcpToken)30 tool callsper day, per IP
Free account token60/day + 1,200 tool callsdaily pacing + rolling 30 days
Member token400/day + 8,000 tool callsdaily pacing + rolling 30 days

To move out of the anonymous bucket, create a token at aigcnews.cn/mcp (no special scopes needed for search) and set it as mcpToken. The token lives in your dsh profile config in plaintext, same as your LLM keys. Daily pacing is checked before the rolling monthly bucket, so hitting the daily limit does not consume monthly quota. When a quota is exhausted the tool returns an actionable error — which bucket, the limit, and how long to wait or where to upgrade — so the agent can relay it instead of failing silently.

Develop from a source checkout

The commands below are for contributors working from a Git checkout. test, verify, and smoke use the built bundle; verify and smoke also call the live MCP endpoint.

pnpm install
pnpm build        # tsdown → dist/
pnpm typecheck    # tsc --noEmit
pnpm test         # node --test (client unit tests against the built bundle)
pnpm verify       # validates registration, cards, routing, and live search
pnpm smoke        # hits the live MCP endpoint through the built client

Load from a dsh source checkout without installing:

pnpm dsh --profile web --patch /path/to/dsh-aigc-radar/cordis.dev.yml

where cordis.dev.yml inserts the row by absolute path:

- insert:
    - id: aigc-radar
      name: /absolute/path/to/dsh-aigc-radar/dist/index.mjs

Data and attribution

Data provided by AIGC Radar — your dsh instance calls the same MCP endpoint that backs the AIGC Radar MCP server, with the same quotas and rate limits. The curated library only covers GitHub projects with 500+ stars; general non-AI GitHub search with direct-GitHub fallback remains an MCP-server capability by design.

License

MIT