Skip to main content

Security report for mcp.so

MCP.so · Technology · N/a

Medium risk. 4 critical or high findings in Security headers, GDPR outside view and 1 other check.

Security score

75/100

Risk levelMedium

0–100, higher is better. How it is calculated

MCP.so is a specialized technology platform offering a comprehensive directory and discovery service for Model Context Protocol (MCP) servers and clients, including integration with Claude MCP. The platform targets developers and AI practitioners seeking to enhance AI capabilities through MCP servers. Technically, the website is built on modern frameworks such as Next.js and React, hosted with Cloudflare DNS and optimized for performance and SEO. Security posture is solid with HTTPS enforced and domain transfer protections, though improvements like DNSSEC and explicit security policies are recommended. Privacy compliance is basic with privacy and terms pages present but lacking cookie consent mechanisms. Overall, MCP.so presents a professional and trustworthy platform with room for enhanced security and privacy features.

Scanned

NIS2 and GDPR posture

GDPR outside view
85/ 100 · Low
Checks completed
7of 7

NIS2 Art. 21(2)

  1. (a) Risk analysis and security policies: Not visible from outside
  2. (b) Incident handling: One indicator
  3. (c) Business continuity, backups and crisis management: Not visible from outside
  4. (d) Supply chain security: Not visible from outside
  5. (e) Secure acquisition, development and maintenance, including vulnerability handling: Partly checked
  6. (f) Assessing the effectiveness of security measures: Not visible from outside
  7. (g) Cyber hygiene and security training: Not visible from outside
  8. (h) Cryptography and encryption: Partly checked
  9. (i) HR security, access control and asset management: One indicator
  10. (j) Multi-factor authentication and secured communications: Not visible from outside

This scan rated NIS2 with an earlier keyword method that is no longer shown; rescan for the per-article view. Risk management, continuity, supply chain, training, MFA, management oversight (Art. 20) and incident reporting (Art. 23) need internal evidence, and Guard cannot tell whether NIS2 applies to this organisation.

GDPR is indicative only, based on publicly visible signals: Articles 7, 12, 13 and 32 of Regulation (EU) 2016/679 and Article 5(3) of the ePrivacy Directive. Not a certification or legal assessment.

Priority findings

4 findings rated critical or high. Fix these first.

  1. Security headers

    Security headers are instructions your web server sends to every visitor’s browser. Missing ones leave visitors more exposed to attacks such as clickjacking and injected scripts.

    • High

      Missing Strict-Transport-Security header

      Forces HTTPS connections

      What to do: Add "Strict-Transport-Security: max-age=31536000; includeSubDomains; preload"

      • CWE-319
      • OWASP A04:2025
      • ASVS V3.4.1
    • High

      Missing Content-Security-Policy header

      Controls resources the browser is allowed to load

      What to do: Implement a Content-Security-Policy appropriate for your application

      • ASVS V3.4.3
      • OWASP Secure Headers: OWASP Secure Headers Project recommends Content-Security-Policy: default-src 'self'; form-action 'self'; base-uri 'self'; object-src 'none';…

  2. GDPR outside view

    These are the privacy signals visitors and data protection authorities look for under GDPR.

    • High

      No Cookie Policy found

      GDPR requires clear information about cookie usage

      What to do: Add a dedicated cookie policy explaining cookie usage and purposes

  3. SSL/TLS

    TLS encrypts the connection and proves the site is really yours. Weak or expiring certificates trigger browser warnings and put traffic at risk.

    • High

      Weak SSL Key Length

      SSL certificate uses 256-bit key, which is considered weak

      What to do: Use at least 2048-bit RSA or 256-bit ECC key

      • CWE-326
      • OWASP A04:2025
      • Mozilla TLS: Mozilla Server Side TLS 6.0 intermediate: rsa_key_size 2048

Need hands-on remediation? Email guard@offseq.com

Keep watching mcp.so

This scan found 18 issues across 6 checks. Monitoring re-runs those checks daily, emails you when something material changes and records each fix, so you can show progress.

Daily monitoring

Recommended

€25per month

  • TLS certificates and expiry
  • DNS records
  • SPF, DKIM and DMARC
  • Security headers
  • GDPR and NIS2 posture
  • Exposed services
  • Third-party trackers
  • Security score drift

Email alerts on material changes and an exportable JSON history of every change.

Billed monthly to your card. Cancel anytime. The full timeline opens with your unlock code.

Full report

€39.90one-time

  • The complete analysis as a PDF you can cite
  • Email delivery with an unlock code
  • Lifetime access to this report

Prices incl. 21% VAT. Secure checkout by Mollie; sold by SEQ SIA, Latvia.

Results by check

Each check with its score, findings and the evidence it collected.

How this score is calculated

The security score runs 0–100, higher is better. It is the plain average of the security checks that completed in this scan, each counted equally. A check that could not run is left out: it counts neither as a pass nor as a fail.

Counted in this score

SSL/TLS
75
Security headers
50
Email security
70
DNS health
70
Network security
100
GDPR outside view
85
Average
75

Not counted

  • NIS2 outside view: Re-weights evidence from the TLS, email, DNS, network and hygiene checks, so counting it would count that evidence twice. Shown on its own.
  • OffSeq Observations: An AI reading of the site, useful context but not reproducible evidence. Shown on its own.

Findings that cap the score at 49

Each of these is serious on its own, so while any is present the site cannot rate better than high risk, however good the other checks are.

  • The certificate is not trusted by browsers
  • The website is not served over HTTPS
  • A database, remote desktop, file-sharing or Telnet service is reachable from the internet
  • The page loads a script from a compromised source
  • The domain registration has expired
  • SPF authorises every server to send mail as the domain

Partial scans

When the site blocks Guard or most checks cannot run, the score covers only the checks that ran (usually DNS and email), is capped at 75 and is shown as partial, without a risk level.

Risk levels

  • Critical0–25
  • High26–49
  • Medium50–75
  • Low76–100

Technical appendix

Business context, site observations and technical notes gathered during the scan, and the standards behind each reference.

Business context

Business profile

MCP.so occupies a niche market within the AI and developer tools ecosystem, focusing on MCP servers and clients. Its business model revolves around community-driven content submission and discovery, supported by advertising revenue via Google Adsense. The platform leverages partnerships with technology providers like SiliconFlow, Tavily, and Cursor to enhance its service offerings. Target customers include AI developers and organizations integrating MCP protocols. The site is relatively new (founded 2024) and small in size but shows consistent branding and a growing ecosystem. The presence of social media and community links indicates active engagement and potential for growth.

Security posture

The website demonstrates a moderate security maturity level with HTTPS enforced and domain registration protections in place. However, the absence of DNSSEC and Content-Security-Policy headers, along with no published incident response or vulnerability disclosure policies, indicates gaps in comprehensive security management. No sensitive data exposure or vulnerable libraries were detected in the provided content. The lack of cookie consent mechanisms despite Google Adsense usage poses a privacy compliance risk. Overall, the security posture is adequate but would benefit from implementing advanced DNS security, CSP headers, and formal security policies.

Recommendations

  1. Enable DNSSEC on the domain to enhance DNS security and prevent spoofing.
  2. Implement a Content-Security-Policy header to mitigate cross-site scripting and related attacks.
  3. Publish a formal security policy and vulnerability disclosure program to improve transparency and incident response readiness.
  4. Add a cookie consent banner/mechanism to comply with GDPR and other privacy regulations due to use of Google Adsense.
  5. Regularly audit third-party scripts and dependencies to ensure no vulnerable libraries are included.
Site observations

Automated reading of the public site. Ratings are indicative.

Business

Company
MCP.so
Description
The largest collection of MCP Servers, including Awesome MCP Servers and Claude MCP integration. Search and discover MCP servers to enhance your AI capabilities.
Key services
MCP server and client directory, Search and discovery of MCP servers, Community submissions and user cases, Integration with Claude MCP and other AI tools
Content quality
good
Branding
consistent

Technology

Stack
React, Next.js, TypeScript, Google Fonts, Google Adsense, Cloudflare DNS
Frameworks
Next.js, React
Platforms
Web
Performance
fast
Mobile
good
Accessibility
basic
SEO
good

Security

Observed score
75/100
Good practices seen
  • HTTPS enforced
  • Domain status clientTransferProhibited
  • No DNSSEC enabled (potential improvement area)

Analytics and tracking

Tracking level
minimal
Privacy practices
basic

Advertising

Ad networks
Google Adsense
Tracking pixels
Google Adsense scripts
Marketing tools
Google Adsense, Funding Choices (Google Consent Management)
Transparency
basic

Site quality

Design
good
User experience
good
Content relevance
good
Navigation
good
Professionalism
good
Trustworthiness
moderate

Key observations

  1. Website is a niche platform focused on MCP servers and clients for AI integration.
  2. Content is well structured with multiple language support (en, zh, ja).
  3. Uses modern web technologies and good SEO practices.
  4. No explicit cookie consent mechanism detected despite Google Adsense usage.
  5. Contact email support@mcp.so is provided but no phone or physical address found.
  6. No explicit security or incident response policies published.
  7. WHOIS data is transparent and consistent with website claims.
  8. Advertising is limited to Google Adsense with basic transparency.
  9. No evidence of adult or questionable content.
Technical notes

Notes

The website is built using a modern React and Next.js stack, leveraging TypeScript for type safety. It uses Google Fonts for typography and integrates Google Adsense for monetization. Hosting appears to be on a performant platform with Cloudflare DNS services. The site is mobile optimized and SEO friendly with comprehensive meta tags and Open Graph data. Accessibility is basic but present. No CMS is detected, indicating a custom or headless approach. The technical implementation is robust with good performance and modern development practices, though security headers could be improved.
Standards and sources

Reference chips under each finding point to the standard, regulation or specification the finding relates to. They are taken from the official datasets below and mapped conservatively: a finding with no defensible link shows none. NIS2 and GDPR references mark where a finding is relevant evidence, not a compliance verdict. Implementing Regulation (EU) 2024/2690 binds digital infrastructure and digital service providers; for other entities it is the most detailed public reading of the Article 21 measures. WCAG, EN 301 549, European Accessibility Act and Web Accessibility Directive references name the requirement an automated accessibility rule relates to; they are not a conformance verdict.

Retrieved 2026-10-25

Reviews of this site

Share your experience with mcp.so. Every review is verified by email and published after moderation.

Overall rating
Select a rating
4000 characters remaining

We’ll email you to confirm your review and keep your details private.

Community rating

No published reviews yet. Be the first to review mcp.so.

Feedback on this report

Rate this report, add an optional comment, and tell us if you want a security specialist to follow up.

Overall report quality
Select a rating
2000 characters remaining

We only use your feedback to improve Guard reports. Contact details are never shared.