The Swiss-Army MCP server for web tools — one install, 15 LLM-ready tools: web search, scraping, screenshots, SEO/a11y audits, DNS/WHOIS/SSL, and more.
ToolCenter MCP demonstrates solid tool design with 15 well-named, action-verb tools covering web operations, DNS, and security checks. All tools have descriptions (avg 180 chars, within baseline 194). Input schemas are present with typed parameters and descriptions. However, output schemas are not documented in the provided source, only inferred from descriptions. Error handling guidance is minimal; tools lack recovery hints and actionable error messages. No tool annotations (readOnlyHint, destructiveHint) despite all being READ_ONLY. Parameter descriptions are adequate but lack format constraints (e.g., URL validation rules, enum values for DNS record types). Missing pagination guidance for tools that may return large result sets.
Run an axe-core accessibility audit on a URL. Reports WCAG violations grouped by impact (critical/serious/moderate/minor) with the rule, description, and a link to the fix guide. Use to check pages before launch or to monitor a11y regressions.
Analyze a URL for SEO issues and return a scored report. Checks title/description length, heading structure, image alt attributes, link health, meta tags, canonical URL, robots.txt, mobile viewport, load time. Returns errors/warnings grouped by severity with recommendations.
Scan a page for broken (4xx/5xx/timeout) links. Returns a summary and the list of broken URLs with their status codes. Great for site maintenance, pre-launch QA, and link-rot monitoring.
Inspect a domain's TLS/SSL certificate — issuer, validity dates, days until expiry, SANs, protocol version, cipher. Flags expired, soon-to-expire (<30 days), or self-signed certs. Use for devops monitoring and pre-launch verification.
HTTP-ping a URL and report whether it's up, its status code, response time, any redirects, and server headers. Use for quick uptime checks and debugging 3xx/4xx/5xx responses.
Output schemas not documented. Tool descriptions infer return structure (e.g., 'returns title, description, Open Graph tags') but no formal schema definition visible in source. LLMs cannot reliably plan downstream tool calls or extract specific fields without documented output types.
No tool annotations despite all tools being READ_ONLY. Tools lack readOnlyHint metadata. This prevents clients from optimizing caching, retry logic, or safety checks. All 15 tools should declare readOnlyHint=true.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 64 | 2026-07-28+ | v2 |
Identify which frameworks, libraries, CMS, analytics, and hosting a website uses (similar to Wappalyzer/BuiltWith). Returns technologies grouped by category with version when detectable. Useful for competitor research and sales prospecting.
Resolve DNS records for a domain. Supports A, AAAA, MX, NS, TXT, CNAME, SOA, SRV, CAA, PTR, or ALL (returns all common record types). Useful for debugging DNS config, finding mail servers, or verifying ownership records.
Extract title, description, Open Graph tags, Twitter Card data, canonical URL, favicon, and author info from a web page. Much faster than `scrape_url` when you only need headline information — perfect for link unfurls or quick context before deciding to read the full page.
Generate a Slack/Discord-style link preview card for a URL — title, description, hero image, favicon, site name, author. Lighter than `get_metadata` and formatted for UI display rather than deep analysis. Useful when an agent needs to show a user what a URL contains.
Fetch a web page and return its main readable content as clean Markdown. Uses Mozilla Readability to strip navigation, footers, ads, and other boilerplate, then converts the article body to Markdown optimized for LLM consumption. Use this instead of `web_search` when you already know the URL and need to read the full content. Executes JavaScript (headless Chrome) so it handles dynamic / SPA pages.
Take a screenshot of a URL and return it as a PNG image. Renders the page in a headless browser and captures the viewport (or full page if fullPage=true). Useful for visual verification, debugging layout issues, or capturing dynamic content.
Convert a web page to a PDF file. Renders the page in a headless browser and exports as PDF. Useful for archiving, printing, or sharing web content in a portable format.
Search the web and return a list of results with title, URL, and snippet. Useful for finding current information, news, or answering factual questions. Supports pagination via limit parameter.
Compare two URLs and return a diff of their content. Useful for detecting changes, monitoring for updates, or comparing versions of a page.
Look up WHOIS information for a domain — registrar, registrant, nameservers, creation/expiry dates, and registration status. Useful for domain research, ownership verification, and expiry monitoring.
Parameter descriptions lack format constraints. E.g., 'url' params state 'Full URL to scrape, must include https://' but no regex pattern or validation rule. 'dns_lookup' type enum is present but other tools lack similar constraints. LLMs cannot validate input before calling.
No error handling guidance. Tool descriptions do not explain what errors are possible, how to recover, or what the LLM should do next. E.g., 'scrape_url' does not document timeout behavior, JavaScript execution failures, or how to handle 404 pages.
Pagination not addressed. Tools like 'web_search' accept 'limit' but no 'offset' or 'cursor' parameter. Tools returning lists (e.g., 'check_broken_links') do not document max result size or pagination strategy. Large result sets risk context window exhaustion.