LucidShark
Local code quality and security scanner: lint, types, SAST, SCA, tests, coverage.
作者提供的说明(英文原文)
Domain security & privacy checker as an MCP server. 17 live checks, 0-100 score, local-first.
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"checker": {
"command": "npx",
"args": [
"-y",
"@oksigenia/checker-mcp"
]
}
}
}在终端中运行这条命令
claude mcp add checker -- npx -y @oksigenia/checker-mcp粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"checker": {
"command": "npx",
"args": [
"-y",
"@oksigenia/checker-mcp"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"checker": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@oksigenia/checker-mcp"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月8日
注册名称已通过域名 oksigenia.com 验证(注册中心会核实域名所有权)
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.1.1)于4个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。
同一分类中有:LucidShark, Threatist, Draugr
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:↓ 每周下载量i★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Local code quality and security scanner: lint, types, SAST, SCA, tests, coverage.
Vulnerability records: CVE, GHSA, CISA KEV, EPSS and vendor fixes in one place. No key needed.
Security scanning for AI agents: SAST, SCA, secrets, IaC, DAST, ranked by real risk.
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
A domain security and privacy checker exposed as an MCP server. It runs entirely on your machine, makes no telemetry calls, and resolves DNS through Cloudflare DoH instead of Google.
Ask any MCP client to scan a domain and get back a scored report covering the certificate, email authentication, DNS and infrastructure, and web security headers, with a plain remediation for anything that fails.
The server exposes three tools.
check_domainScan a domain or URL and return a scored report from real live checks.
// arguments
{ "domain": "example.com", // bare domain or full URL "categories": ["ssl", "email"], // optional subset; defaults to all "format": "report", // "report" (default) or "json" "lang": "en" // "es" or "en" }
list_checksList every check grouped by category, with the reason each one matters.
{ "lang": "en" }
explain_checkExplain a single check by id, why it matters, and how to remediate a failure.
{ "check_id": "dmarc", "lang": "en" }
Seventeen checks across four categories, weighted to a 0-100 score.
| Category | Check id | What it verifies |
|---|---|---|
| ssl | ssl | SSL/TLS certificate validity and expiry |
spf | SPF record present and not overly permissive | |
dmarc | DMARC policy and reporting | |
dkim | DKIM signing (common selectors probed) | |
mta_sts | MTA-STS against SMTP downgrade attacks | |
mx | MX records for mail delivery | |
bimi | BIMI brand logo record and VMC | |
| dns | dnssec | DNSSEC signing on the zone |
| dns | ptr | PTR / forward-confirmed reverse DNS |
| dns | caa | CAA records restricting certificate issuance |
| dns | blacklist | Malware/phishing reputation |
| dns | expiry | Domain registration expiry |
| dns | whois_redact | WHOIS/RDAP personal-data redaction |
| web | headers | HTTP security headers (HSTS, CSP, X-Frame-Options, …) |
| web | csp | Content-Security-Policy quality |
| web | web_tech | CMS/server version and API exposure |
| web | hsts_preload | HSTS preload-list status |
Run list_checks for the full descriptions, or explain_check for one at a time.
Run it straight from npm with npx:
npx @oksigenia/checker-mcp
Register it with Claude Code:
claude mcp add oksigenia-checker -- npx -y @oksigenia/checker-mcp
Or add it to any MCP client config (works in Claude Desktop, Claude Code, or any MCP client):
{
"mcpServers": { "oksigenia-checker": { "command": "npx", "args": ["-y", "@oksigenia/checker-mcp"], "env": { "OKSIGENIA_LANG": "en" } } } }
Then ask: *"Check the security of example.com"* or *"Explain the dmarc check."*
Build and run it as a container. It speaks MCP over stdio, so run it attached to your client's stdin/stdout (-i), not on a port.
docker build -t oksigenia-checker-mcp .
docker run --rm -i oksigenia-checker-mcp
Reports render in Spanish and English. Set the default with OKSIGENIA_LANG (es or en), or pass lang on any tool call. It defaults to English.
The only thing that leaves your machine is the domain name you scan, sent to:
cloudflare-dns.com) and Cloudflare's security resolver, for DNS and reputationrdap.org), for registration and WHOIS-privacy dataThere is no telemetry, no analytics, and no third-party scoring service. Nothing about your queries is stored or transmitted anywhere else.
npm install
npm run build # tsup -> dist/ npm run typecheck # tsc --noEmit npm test # vitest (hermetic, mocked network) npm run smoke # spawn the built server and exercise every tool live
Dev watch: npm run dev.
src/engine.ts — the network engine. Language-agnostic: returns i18n keys, not text. Every check is individually resilient and runChecks never throws.src/i18n.ts — es/en dictionaries and t().src/index.ts — the MCP server: renders engine keys into the requested language and exposes the three tools.See CONTRIBUTING.md for the PR flow.
GPL-3.0-or-later © Oksigenia SL