TentacleTool
触手怪的 MCP 工具集(monorepo)— 一个文件夹 = 一个独立工具集,全部零依赖 Node 原生实现,走 MCP Streamable HTTP 协议各自监听端口
Install / Use
claude mcp add yxpil -- npx -y github:yxpil/TentacleToolIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Our assessment of TentacleTool
TentacleTool scores 73/100 on our quality scale, 3734th of 4,588 Development & Engineering skills we index.
Its MCP Server is 3.8 KB long, well organised into 9 sections with 2 code examples: a solid amount of guidance for an agent.
It has 3 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated 2 days ago, so TentacleTool is actively maintained.
- It is released under the Apache-2.0 license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 87/100, with 2 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-10-08. Automated pattern scan on 2026-10-08. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
TentacleTool compared with similar skills
All 4 of these similar skills score higher than TentacleTool; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| TentacleTool (this skill)by yxpil | 73 | 3 | 2d ago | MCP Server |
| Agent-Reachby Panniantong | 100 | 93.2k | today | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 74.6k | today | CLAUDE.md |
| CowAgentby zhayujie | 100 | 47.3k | today | CLAUDE.md |
| ai-job-searchby MadsLorentzen | 100 | 45.2k | 2d ago | CLAUDE.md |
Frequently asked questions
- How do I install TentacleTool?
- Run
claude mcp add yxpil -- npx -y github:yxpil/TentacleTool. The install tabs above show the steps for each supported agent. - Which AI agents does TentacleTool work with?
- It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
- Is TentacleTool safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is Apache-2.0-licensed and scores 87/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is TentacleTool still maintained?
- The repository was last updated 2 days ago, so TentacleTool is actively maintained.
Skill content
View source on GitHubTentacleTool
工具集单仓库(monorepo):一个文件夹 = 一个独立工具集,互不依赖,可单独启动。
每个工具集都是一个零依赖 MCP 服务器(只依赖 Node 原生模块,不装任何 npm 包), 走 MCP Streamable HTTP 协议(2025-03-26),本地监听一个固定端口。
工具集一览
| 文件夹 | 工具集 | 端口 | 工具数 | 说明 |
|--------|--------|------|--------|------|
| neton/ | NetON | 8341 | 6 | 局域网网络工具(设备发现 / 端口扫描 / 协议分析 / 抓包) |
| webview/ | WebView | 8342 | 4 | 网页转 Markdown(网页转 MD / HTML 转 MD / 链接提取 / 元信息提取) |
| search/ | Search | 8343 | 3 | 聚合搜索(必应/百度/DuckDuckGo,折叠摘要 + 翻页 + 按需展开全文) |
| find/ | Find | 8344 | 4 | 本机文件搜索(Everything 风格:文件名 / 最近修改 / 内容检索 / 工具定位,可选 C 原生索引器加速) |
| calc/ | Calc | 8345 | 5 | 科学计算器(表达式求值 / 单位换算 / 方程与方程组 / 矩阵运算,手写零依赖引擎) |
| analyze/ | Analyze | 8346 | 8 | 代码知识图谱(符号搜索 / 引用 / 调用链 / 依赖与循环检测 / 最短路径 / 影响面分析,11 种语言零依赖提取器) |
| kb/ | KB | 8347 | 6 | MySQL 知识库(把多张表组合成一个有业务含义的知识库:跨库跨源组合 / 跨表全文搜索 / 只读 SQL / 结构内省,零依赖自实现 MySQL 客户端协议) |
| stamp/ | Stamp | 8348 | 6 | 时间与调度(时间戳转换 / 时长解析 / 时区换算含 DST / 工作日推算含调休 / cron 解析与触发预测,纯 Intl 实现时区运算) |
| jsonx/ | Jsonx | 8349 | 6 | 数据格式处理(JSON/YAML/CSV/TSV 互转含往返校验 / JSONPath 查询 / 结构推断 / 结构化 diff / 统计聚合,CSV 走完整状态机、类型推断保守化) |
| fsx/ | Fsx | 8350 | 10 | 文件系统操作(读 / 写 / 精确替换 / 列目录 / 目录树 / 属性 / 内容检索 / 复制 / 移动 / 带四道安全闸门的删除) |
| gitx/ | Gitx | 8351 | 10 | Git 仓库操作(状态 / 历史 / 差异 / 逐行归属 / 文件历史 / 分支 / 暂存 / 远程 / 提交,刻意不做 push 与任何改写历史的操作) |
| httpx/ | Httpx | 8352 | 6 | HTTP 客户端(通用请求 / 下载 / 头探测 / 并发批量 / JSON API / 站点探测,含 SSRF 防护与内网拦截) |
| cryptox/ | Cryptox | 8353 | 9 | 哈希 / 编码 / 加密(摘要 / HMAC / 文件校验和 / 编解码 / JWT 解析验签 / UUID 与短 ID / 密码生成与强度评估 / AES-256-GCM 加解密,全程本地无网络) |
下一个新工具集用 8354。测试端口约定:18000 + 正式端口(如 8348 → 18348、8353 → 18353)。
快速开始
每个工具集自带独立的 package.json、README.md 和 start.bat。两种启动方式:
# 方式一:Node 直接跑(推荐,能看到实时日志)
cd kb
node src/index.js
# 方式二:双击或调用 start.bat(后台最小化启动,并自检端口)
自定义端口用环境变量覆盖,例如 $env:KB_PORT=9347; node src/index.js。
接入 MCP 客户端
服务器起来后,在 MCP 客户端里按「Streamable HTTP」类型添加,URL 形如:
http://127.0.0.1:8347/
连上后先调该工具集的「入口工具」了解可用能力(例如 kb_sources、analyze_stats,
或 calc_help),再按需调用具体工具。
测试
每个工具集自带零依赖测试,直接 node test/<某测试>.js,也可在文件夹内 npm test:
| 文件夹 | 单测 | 端到端 |
|--------|------|--------|
| neton | npm test(80 项) | — |
| webview | npm test(100 项) | — |
| search | npm test(62 项) | — |
| find | npm test(74 项) | — |
| calc | npm test(59 项) | node test/mcp.e2e.js(28 项) |
| analyze | npm test(23 项) | node test/mcp.e2e.js(57 项) |
| kb | npm test(34 项) | node test/mcp.e2e.js(60 项) |
| stamp | npm test(351 项) | node test/mcp.e2e.js(58 项) |
| jsonx | npm test(333 项) | node test/mcp.e2e.js(159 项) |
| fsx | npm test(88 项) | node test/mcp.e2e.js(51 项) |
| gitx | npm test(75 项) | node test/mcp.e2e.js(54 项) |
| httpx | npm test(50 项) | node test/mcp.e2e.js(65 项) |
| cryptox | npm test(159 项) | node test/mcp.e2e.js(66 项) |
合计 单测 1488 项 + 端到端 598 项,全部通过、无跳过。
单测是纯逻辑层(不联网、不扫盘),可随时回归;端到端会自起服务器走完整 JSON-RPC。 需要外部依赖的用例(数据库、外网)在依赖不可用时会优雅跳过而不是失败。
新增工具集
- 在根目录新建一个文件夹(与其他工具集平级)
- 文件夹内自带
README.md(用途与启动方式)、package.json、start.bat、test/ - 更新本文件的「工具集一览」表格与端口
通用约定
- 零依赖:只用 Node 原生模块;确需加速时可自带预编译的本地组件(如
find/的 C 索引器) - 上下文经济:工具输出默认折叠 + 限制条数 + 给翻页/续读提示,不把整页内容灌回模型
- 失败要可诊断:错误信息说清「哪一步错了 + 怎么改」,而不是抛原始异常
.gitignore作用于整个仓库(node_modules/、logs/、*.log、.env、cache/)- 凭据类文件一律不入库:如
kb/kb.config.json(数据库连接串含账号密码), 仓库里只放*.example.json模板
<div align="center"> <a href="https://github.com/yxpil/TentacleTool"> <img width="100%" src="https://alittlecatgirlpanel.yxp.hk/card?repo=yxpil/TentacleTool" alt="gh-card · yxpil/TentacleTool" /> </a>
<sub>Powered by <a href="https://alittlecatgirlpanel.yxp.hk"><b>gh-card</b></a> · 粉色手写体 README 仓库名片</sub>
</div>Related Skills
Agent-Reach
93.2kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
74.6kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
CowAgent
47.3kOpen-source personal AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-agent, multi-model, multi-channel. Lightweight, extensible, one-line install.
ai-job-search
45.2kThe job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
