Skip to main content
Glama
hyunwk

oss-trends-mcp

by hyunwk

Oops typos "贪腐" etc. Need careful.

Let me carefully di-re:

Title: "# oss-trends-mcp"

Para1: "面向 AI 开发者的任务驱动型热门 OSS 推荐器,以 MCP 服务器形式对外提供。只要输入“构建 X 时,有哪些 OSS 正在流行?”送到即可获得由实时信号支撑的开源项目排名:GitHub star 增长动能、npm/PyPI 下载增长,以及 Hacker News 热度。"

"只要咨询" or "询问". Better: 询问"构建 X 时,什么 OSS 正在流行?"即可得到……

Para2: "Awesome 列表几周就会过时;AI 工具生态则以天为单位变化。oss-trends-mcp 在提问时刻基于真实信号而不是手工精选的链接对候选排序——将成熟的玩家与潜在的上星新人(repo 有18 个月的仓库)区分开。"

Ok.

Now description of recommend_oss: "Returns a permanent ranking plus newcompers" - "返回 established 榜单和 rising 新人,每条都带揭示性评分 breakdown".

Let me write official:

recommend_oss 主工具

"输入:task(你想构建什么)、queries(适配的分 GitHub 搜索词——;否则从任务文本中自动扩展)、ecosystem(auto/npm/pypi/none)、limit。返回 established 排名列表以及 risingNewcomers 两列,每个结果 get a transparent score breakdown, with a LLM can explain "why". LLM 负责将 task 展开为搜索词;服务器负责确定性的收集与打分。"

I'm mixing too much English. That's okay because those terms are untranslated. But needs enough Chinese.

Let's final:

"输入:task(你想构建的内容)、queries(有重点的 GitHub 搜索词——建议指定;否则自动从任务文本扩展)、ecosystem(auto/npm/pypi/none)、limit。输出为一个 established 排名,外加 risingNewcomers,各自附上透明的评分明细,便于调用方 LLM 解释“为什么”。任务展开为搜索词的动作由 LLM 完成;服务器进行确定性的收集和评分。"

"提供 GitHub 仓库原始搜索接口,输出规范化的结果。"

repo_health

"单个仓库深度信号:star 动能、维护时效性、npm/PyPI 采用、HN 热议度,还有 0-100 分数。"

compare_repos

"2-6 个仓库逐项趋势比较,按照综合得分排序。"

Scoring:

"## 评分(0-100)"

Bullets:

  • momentum 35 — 近 30 日拣获的 star 数量(log 刻度)

  • maintenance 20 — push 时效的衰减;归档仓库 0 分

  • adoption 20 — npm 与 PyPI 月下载量+相对最近 28 天与前 28 天的增长奖励

  • community 15 — 网络扶持:forks、license、topics

  • buzz 10 — 最近 180 天产生的 Hacker News 故事和其中的最高得分

Setup:

"构建并注册到 Claude Code:GXP1 如果给服务器提供 GITHUB_TOKEN,否则每分钟只能 10 searches 和每小时 60 core requests,这样采样 star 动量的星星会多降级为 fallback:

GXP2 能在任何支持 stdio MCP 客户端(Claude Code,Cursor 等)中运行"

Need "fallback signals" translated.

Ok. Let me avoid "GPLO".

Now final output including exact bullet marks with em dash — preserve.

Let me get precise. I'll now write the final variant carefully.

I also need to ensure "10 searches/min and 60 core requests/hour" is translated "每分钟 10 次搜索和每小时 60 次核心请求" and may keep as "10 次搜索/分钟和 60 次请求/时". GITHUB basic:allow unauthenticated rate limits 10 search/min, 60 core/hour.

"因此 star 动能采样常常只能回退到其他信号" - ok.

For "unmatched repos skip adoption signals" -> "未匹配的仓库会省略 adoption 信号" good.

"the calling LLM" -> "调用方 LLM".

"transparent score breakdown" -> "透明的得分拆解".

"deterministic collection and scoring" -> "确定性的收集与打分".

"normal placement output" - "标准化的输出"

Now to "article skeleton": no.

Let's produce final.

One more check: "Awesome list" might be a given name; writing "Awesome 列表" keeps "Awesome" word. Good.

Also "AI builder" maybe "AI 构建者" or "AI 构建方向的人"; I'll use "AI 开发者" as common.

"task-driven" - "任务驱动" is a readable term.

Now the close. Let me craft the entire final in one shot:

A
license - permissive license
Not graded
quality - not tested
C
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    Not graded
    quality
    C
    maintenance
    Discover, rank, and compare GitHub repositories from any MCP-compatible AI client. Enables searching, filtering, ranking, and evaluating open-source repositories by topic, language, stars, license, activity, and relevance.
    MIT
  • F
    license
    Not graded
    quality
    C
    maintenance
    Provides AI agents with 6 tools for searching Hugging Face models, GitHub trending repos, analyzing GitHub repositories, fetching dev.to articles, Show HN launches, and Product Hunt daily launches. Built for dev tooling research and AI ecosystem analysis.
  • A
    license
    Not graded
    quality
    F
    maintenance
    Open source intelligence for AI agents — GitHub project-health scoring, package dependency-risk analysis, trending repositories, license checks, and side-by-side package comparison.
    MIT

View all related MCP servers

Related MCP Connectors

  • GitHub project health, package dependency risk, trending repos, license & package comparison.

  • npm, PyPI & crates.io intel for AI agents: versions, popularity, deps, health. No API keys.

  • Evidence-backed open-source project search, recommendations, alternatives, and comparisons.

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/hyunwk/oss-trends-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server