nemiget.

Security check

We look at your site the way attackers do.

Scanners try every site on the internet for open .env files, git folders, admin panels and weak passwords. Our own servers block these requests every day; the live counter is just below. We check your website, web app or API for the same holes and for the ones scanners can't see, then give you a report with exact steps to close them.

Blocked in 24 h —

Methodology

  • OWASP WSTG
  • OWASP ASVS
  • OWASP Top 10
  • API Security Top 10
  • CWE
  • CVSS 3.1

We test against the open OWASP standards instead of a home-made method. Every finding comes with a CWE number and a CVSS score, so your developers know right away what it is about. We don't call it a certified audit: we hold no security certifications, and we say so.

Two levels

Choose how deep we go.

We don't need access to your code or servers: we look at things the way an attacker would. Every check starts with your written permission.

External2–3 working days

Your site from the outside

Free for the first 5 companies, in exchange for a review
What we see
What any visitor or scanner sees: domains, servers and ports, exposed files and admin panels, HTTPS and browser protection, forms, public API methods.
Standard
The part of the OWASP ASVS level 1 requirements that can be checked from the outside.
What you give us
The list of addresses and a signed permission.
With accounts4–6 working days

If you give us test accounts

What we see
Everything from the outside, plus what you see after logging in: access to other users' data, role permissions, the API behind login, order and payment logic.
Standard
OWASP ASVS level 2: the key sections that can be checked through the app.
What you give us
Test accounts for each role, ideally on a staging copy. If you'd rather not, we check from the outside only.

What we check

The holes we look for.

Exposed files

.env with keys, .git, backups and database dumps, debug pages, source maps.

Admin panels and ports

phpMyAdmin, Grafana, hosting and CMS panels open to the whole internet; databases, Redis and Docker on open ports; cameras and routers on your addresses.

Forgotten parts

Old subdomains, test copies of the site, subdomains that anyone can take over.

Cloud and storage

Open S3, Google Cloud and Azure buckets, public backups, cloud keys in page code. Only what is visible from the outside.

Login and sessions

Password guessing limits, password reset, two-factor bypass, JWT and OAuth mistakes.

Other people's data

Can one user open another user's orders, files or profile by changing a number. Can a user act as admin.

API: REST and GraphQL

Per the OWASP API Top 10: other users' objects by ID, excess data in responses, admin-only methods, page-by-page database scraping, open docs and old versions.

Injections

SQL and NoSQL injection, XSS, server-side requests to your internal network, file uploads that run code.

Front-end scripts

DOM XSS, third-party scripts that can be swapped, keys left in JavaScript, tokens in browser storage.

WebSocket

Can another site connect on behalf of your user, are rights checked on every message, can one user join another's channel.

Business logic

Negative quantities, a promo code used twice in parallel, a fake “paid” callback, skipped order steps.

Bots and abuse

Whether a simple script posing as a browser gets through login, sign-up and checkout. Rate limits, form spam, bonus and promo code abuse, data scraping.

AI chatbots

Prompt injection, leaks of other users' documents through search, tools the bot should not call, burning your model budget.

Email and domain

Can someone send email in your name (SPF, DKIM, DMARC), certificate and DNS settings.

HTTPS, headers, versions

Encryption, security headers, cookies and CORS, outdated server, CMS and library versions with known vulnerabilities.

Rules we keep

Careful with your live site.

We do

  • We work from the outside: we don't need access to your code, servers or cloud.
  • Work only with your written permission, an NDA and proof that the domain is yours.
  • Keep a log of every request we send, so you can tell us apart from a real attack.
  • Call you the same day if we find something that can be used right now.
  • Prove a problem with one masked record and stop there.
  • Delete test accounts, tokens and your data when the work is done.

We never

  • Run load tests or anything that can take your site down.
  • Phish your employees or call them pretending to be someone else.
  • Log in to real customers' accounts or guess real passwords.
  • Download, change or delete your data.
  • Test services you don't own: payment providers, messengers, cloud hosts.

What you get

A report your team can act on.

In the report

  • One page for management: how serious it is and what to fix first.
  • For every finding: where it is, what an attacker could do in business terms, steps to reproduce, proof, and how to fix it in your stack.
  • Every finding has a CWE number, an OWASP Top 10 category and a CVSS score.
  • An OWASP ASVS checklist: which requirements we checked and with what result.
  • What we checked and found fine, so you see the full picture.
  • A fix plan: today, this week, this month.
  • In English, Russian or Spanish. PDF plus a table for developers.
  • One re-check within 30 days after you fix. The report gets a second version with each finding marked fixed or not.

How we rate findings

By the real risk to your business, with the CVSS score next to it.

CriticalOpen .env with keys, admin without a password, other customers' orders by number. Fix today.
HighStored XSS, fake payment callback, no login limit. Fix this week.
MediumEmail in your name can be spoofed, outdated software, detailed error pages. Fix this month.
LowSmall gaps worth closing with the next release.
NoteAdvice and things done well.

A check lowers the risk. No one can promise that a site has no holes at all, and we don't.

Scope

What the check doesn't cover.

We check only what is listed in the permission, and only as of the date of the check. These limits are written in the permission you sign and in every report.

We don't check

  • Your office's internal network: Wi-Fi, employees' computers, cameras and devices inside the network.
  • Physical access and your staff: entrances, phishing, phone calls.
  • Services you don't own: payment providers, messengers, cloud and hosting.
  • Addresses, accounts and apps that are not in the permission.
  • Your code, servers from the inside and cloud accounts: we don't ask for access to them.

Good to know

  • The report shows the state as of the date of the check. New code, new services and new people with access are not covered.
  • Cameras and devices visible from the internet on your addresses go into the report, but we don't log in to them.
  • You can fix what we find yourself or order the fixes from us.

How it works

From the first call to the re-check.

  1. Call

    30 minutes: what the project is, what matters most to protect, which level fits.

  2. Paperwork

    Permission, NDA, a DNS record or file that proves the domain is yours. Test accounts, if you decide to give them.

  3. Check

    In the time window you approve. Critical issues reported the same day.

  4. Report

    Within 2 working days after the check, then a call with your developers.

  5. Re-check

    After your fixes. Then we delete your data and confirm it in writing.

Contact

Tell us what to check.

Send the address of your site or app and what worries you most. The founder replies personally, usually within a day. The estimate is free.

Nemiget · engineering studio · nemiget.com