Skip to content

enterno-io/enterno-website-diagnostics

v1.0.3MIT

Website and domain diagnostics for AI agents: HTTP, redirects, SSL/TLS, DNS, MX, DKIM, WHOIS, DNS propagation, CSP, CORS, cookies, robots.txt and structured data. 19 checks with no signup; a free Enterno.io key adds 40+ MCP tools (SEO audit, AI-readiness, security grade).


The enterno-website-diagnostics agent skill teaches Claude Code, Codex, Cursor and other agents how to check a public website or domain and how to turn the results into a report a person can act on. The checks run on Enterno.io servers, so the site is seen the way the public internet sees it — not from your laptop or VPN.

It works the moment it is installed: 19 checks need no account and no key. Connect the hosted Enterno.io MCP server with a free key and the agent gets 40+ tools, including an on-page SEO audit, AI-readiness, a full security grade from A+ to F and email deliverability.

Contents

Install

Any agent supported by the skills CLI (Claude Code, Codex, Cursor, Gemini CLI and others):

npx skills add enterno-io/enterno-mcp --skill enterno-website-diagnostics

Claude Code plugin — the repository is also a plugin marketplace:

/plugin marketplace add enterno-io/enterno-mcp
/plugin install enterno-website-diagnostics@enterno-io

Codex, Gemini CLI, GitHub Copilot CLI and Cursor — the same repository is a plugin for them too (Agent Plugins standard and a Gemini CLI extension):

# Codex
codex plugin marketplace add enterno-io/enterno-mcp
codex plugin add enterno-website-diagnostics@enterno-io

# Gemini CLI
gemini extensions install https://github.com/enterno-io/enterno-mcp

# GitHub Copilot CLI
copilot plugin marketplace add enterno-io/enterno-mcp
copilot plugin install enterno-website-diagnostics@enterno-io

# Cursor
npx skills add enterno-io/enterno-mcp --skill enterno-website-diagnostics -a cursor

By hand — copy skills/enterno-website-diagnostics into your agent's skills folder, for example ~/.claude/skills/ for Claude Code.

No configuration is needed for the 19 open checks. For the full toolset, add the MCP server — see Connect the MCP server.

What you can ask

  • "Is example.com down? Find which layer fails: DNS, TLS, HTTP or redirects."
  • "Run a pre-launch checklist for shop.example.com."
  • "We moved DNS yesterday — has every resolver picked up the change?"
  • "When do the certificates of these five domains expire, and how good is their TLS setup?"
  • "Grade the security headers and cookies of my site and tell me exactly what to add."
  • "Why do our emails go to spam? Check MX, SPF, DKIM and DMARC."
  • "Is my robots.txt blocking AI crawlers? Is the structured data valid?"

See it in action

Check when the certificates of github.com, wikipedia.org and example.com expire and how good their TLS setup is. Then grade github.com's security headers and cookies.

Real run, 11 November 2026, open API mode, 5 requests:

Verdict: all three certificates are valid for at least 44 more days; example.com has a weaker TLS setup than the other two.

SeverityFindingEvidenceHow to fix
warningexample.com still accepts TLS 1.0 and 1.1 and sends no HSTSTLS grade B (80/100): "TLS 1.2 (legacy enabled)" 10/20, "No HSTS" 0/10Disable TLS 1.0/1.1 at the edge; add Strict-Transport-Security: max-age=31536000; includeSubDomains
infoexample.com certificate renews soonestCloudflare TLS Issuing ECC CA 3, valid to 2026-12-26, 44 days leftNothing to do if auto-renewal is on; re-check under 30 days
passgithub.com TLSSectigo DV E36, valid to 2027-01-28 (77 days), grade A+ (97), TLS 1.2/1.3 only, HSTS strong—
passwikipedia.org TLSLet's Encrypt YE2, valid to 2027-01-02 (52 days), grade A+ (97), TLS 1.2/1.3 only—
infogithub.com _octo cookie readable by scriptscookie grade B: _octo Secure, SameSite=Lax, no HttpOnly; _gh_sess and logged_in have all three flagsAdd HttpOnly if no script needs to read it

Security headers grade for github.com: B (80/100), HSTS and HTTPS full marks. Not checked: SEO audit and AI-readiness (need the free MCP connection).

What it checks

AreaNo key (open API)With a free key (MCP server)
HTTPstatus, response headers, redirect chain, HTTP/2 and HTTP/3, short-link expansion+ mixed content, final URL resolution, Wayback history
TLScertificate, chain, SAN, expiry, protocols, TLS gradesame
DNS and domainsrecords, propagation across public resolvers, MX, DKIM, WHOIS, ASN, IP data+ SPF/DKIM/DMARC in one check, reverse IP, subdomains from CT logs
Securitysecurity-headers grade (from headers), Content-Security-Policy, CORS, cookie flags+ full security grade A+ to F
EmailMX with SMTP banner, DKIM+ full audit (SPF, DMARC, BIMI, MTA-STS), STARTTLS test, raw header parsing
SEO and AIrobots.txt with AI-crawler policy, structured data, Open Graph+ on-page SEO audit, AI-readiness grade, technology and CMS detection, PageSpeed
Networkping+ traceroute
Your account—monitors (list, create, pause), observed assets and events

Full tool list: TOOLS.md. Open API reference: open-api.md.

Built-in recipes

The skill carries step-by-step playbooks, so the agent runs the right checks in the right order:

  • "Is the site down?" — walks HTTP → DNS → ping → TLS and names the first layer that fails, with evidence.
  • Pre-launch / post-migration checklist — DNS, propagation, certificate coverage of www and apex, redirects, security and cache headers, HTTP/2 and HTTP/3, robots.txt, structured data.
  • Security headers and cookies — what is missing, why it matters, the exact header to add.
  • Email deliverability — MX with SMTP banner, SPF and DMARC from DNS, DKIM by selector; the full audit through the MCP server.
  • Certificate expiry across many domains — one table, soonest first.

Every report follows the same shape: a one-line verdict, a findings table (severity, finding, evidence, fix) and an honest list of what was not checked.

Connect the MCP server

The skill uses the MCP server automatically when its tools are available.

Endpointhttps://enterno.io/mcp
TransportStreamable HTTP (JSON-RPC 2.0), protocol version 2024-11-05
AuthenticationAPI key: Authorization: Bearer ent_… or X-API-Key: ent_…
API keyFree — create an account at https://enterno.io/en/register, then create a key in the dashboard
Tools49 tools (server card, 2026-11-11) — full list in TOOLS.md
Discovery/.well-known/mcp.json, /.well-known/mcp/server-card.json
Registryio.enterno/site-diagnostics in the official MCP Registry
Documentationhttps://enterno.io/en/mcp-server

The initialize handshake is open. tools/list and tools/call need a key; without one the server answers HTTP 401 API key required. Keep the key out of files you commit: use an environment variable or the client's secret prompt. Configs below are also in examples/.

Claude Code

read -rs ENTERNO_API_KEY   # paste the key; it is not echoed or saved to shell history
claude mcp add --transport http enterno https://enterno.io/mcp \
  --header "Authorization: Bearer $ENTERNO_API_KEY"

This stores the key in your user-level Claude Code config. Do not add it with --scope project: that writes .mcp.json into the repository, where it gets committed.

Claude Desktop

Custom connectors in Claude Desktop sign in with OAuth and cannot send an API-key header, so connect through the open-source mcp-remote bridge (Node.js 18+). In claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):

{
  "mcpServers": {
    "enterno": {
      "command": "npx",
      "args": ["-y", "mcp-remote@0.15.1", "https://enterno.io/mcp", "--header", "Authorization:${ENTERNO_AUTH}"],
      "env": { "ENTERNO_AUTH": "Bearer ent_your_key" }
    }
  }
}

Restart Claude Desktop. The key sits in plain text in this user-level file, so keep the file private. mcp-remote is a third-party package that sees your key; the config pins a version you can review — update it deliberately. Its README also documents --header-file, which reads the header from a separate file instead of the config.

Codex CLI

~/.codex/config.toml — Codex reads the token from the environment (export ENTERNO_API_KEY=ent_...):

[mcp_servers.enterno]
url = "https://enterno.io/mcp"
bearer_token_env_var = "ENTERNO_API_KEY"

Cursor

~/.cursor/mcp.json (global) or .cursor/mcp.json (project). Cursor reads ${env:NAME} from your environment:

{
  "mcpServers": {
    "enterno": {
      "url": "https://enterno.io/mcp",
      "headers": { "Authorization": "Bearer ${env:ENTERNO_API_KEY}" }
    }
  }
}

VS Code

.vscode/mcp.json — VS Code asks for the key once and stores it as a secret:

{
  "inputs": [
    { "type": "promptString", "id": "enterno-api-key", "description": "Enterno.io API key (ent_…)", "password": true }
  ],
  "servers": {
    "enterno": {
      "type": "http",
      "url": "https://enterno.io/mcp",
      "headers": { "Authorization": "Bearer ${input:enterno-api-key}" }
    }
  }
}

Zed

Your user settings.json (~/.config/zed/settings.json) → context_servers. Never put the key in a project's .zed/settings.json — that file is usually committed.

{
  "context_servers": {
    "enterno": {
      "url": "https://enterno.io/mcp",
      "headers": { "Authorization": "Bearer ent_your_key" }
    }
  }
}

Any client / curl

curl -s https://enterno.io/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $ENTERNO_API_KEY" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"check_ssl","arguments":{"host":"example.com"}}}'

More in examples/curl.sh.

Plans and limits

  • Open API — no key, per tool and per IP, 10 to 30 requests a minute. A full audit of one site takes about 15 requests. Plain text, JSON or CSV: https://enterno.io/api/open/<tool>?q=<target> (docs).
  • MCP server — the key is free. Tool calls count against your plan's daily API allowance; the numbers per plan are on https://enterno.io/en/pricing. Every tool works on a free key: the plan changes how many calls a day you get, not which tools you can call.
  • Rate limiting is per key (sliding window); responses carry X-RateLimit-* headers.

Security and privacy

  • Only the hosts you name are checked. The skill does not crawl beyond them or probe ports and subdomains on its own.
  • Requests to private, loopback and internal addresses are refused by the server.
  • The skill never asks you to paste a key into the chat and never writes it into project files.
  • No local files are read or sent, and the skill ships no scripts — only instructions and reference docs. Config samples for each client are in examples/.

FAQ

Do I need an account? Not for the 19 open checks. A free account gives the API key for the MCP tools.

Which agents does it work with? Any agent that can make an HTTP request or use MCP tools: Claude Code, Claude Desktop, Codex CLI, Cursor, VS Code, Zed, Gemini CLI and others.

Where do the checks run? On Enterno.io servers, so results do not depend on your network, VPN or DNS cache.

What language does the report come in? The language you ask in.

Is the MCP server open source? No. This repository holds the skill, the documentation and the client configs; the server is a hosted service run by enterno.io with a free tier.

Keeping TOOLS.md current

TOOLS.md is generated from the public server card, not written by hand:

scripts/update-tools.sh   # needs bash, curl, jq

Support

Questions and bug reports: open an issue or write via https://enterno.io/en/contact.

License

The skill, documentation and examples in this repository are MIT-licensed — see LICENSE. The Enterno.io service is proprietary and has a free tier.