Techbanq

Incident response

Do not restore from backup yet

Restoring first feels like the fastest route back to normal. It overwrites the evidence of how they got in, and if the entry point is still there, it puts the same site back on the same hole.

Serving Saudi Arabia from our Riyadh office and our engineering hubs in Lahore and Karachi

Workstation showing terminal output during a security review

Enterprise delivery credentials

  • National Cybersecurity Authority, Saudi Arabia

    NCA registration

    Registered cybersecurity service provider

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

First hours

The order that keeps your options open

Read this before doing anything irreversible. Most of it takes minutes and costs nothing, and each step preserves something the next one would have destroyed.

STAGE 01

1 · Take an image before you touch anything

A snapshot of the server or container, and a copy of the web root, the database and every log you have. Store them somewhere separate and do not work on the originals. This is the step people skip, and it is the one that cannot be recovered later. The evidence still exists after the cleanup.

STAGE 02

2 · Contain without destroying

Take the site offline or behind a maintenance page rather than deleting files. Rotate the credentials the site uses — database, mail, API keys, hosting panel — and revoke sessions. Do not reinstall yet. The attacker loses access; the evidence survives.

STAGE 03

3 · Establish what was reachable

What the compromised account or process could read: the customer database, the payment integration, the mail relay, other sites on the same host. This determines whether the incident is a defacement or a data breach, and the two carry very different obligations. A scope statement your legal and compliance function can act on.

STAGE 04

4 · Find the entry point before rebuilding

Access logs around the first sign of change, file modification times, and any account that authenticated from somewhere unusual. Rebuilding without this puts the same application back on the same defect. A named cause, or an explicit statement that the cause is unknown.

STAGE 05

5 · Rebuild rather than clean

A fresh host, the application deployed from source control, the data restored from a backup predating the compromise, and the entry point closed before it goes back online. Removing the files you found leaves the ones you did not. A site whose state you can account for, rather than one that has gone quiet.

How they got in

Where the entry point usually turns out to be

  1. 01

    An out-of-date plugin or component

    The most common cause by a wide margin, and it is rarely the core platform — it is an extension installed for one feature three years ago, still active, still loaded on every request, and unmaintained upstream since.

  2. 02

    A credential that was reused

    An administrator password that also appears in a public breach corpus, or a hosting panel login shared with a former contractor. No exploitation involved; somebody simply logged in, and the access log will show it.

  3. 03

    A neighbour on shared hosting

    Another site on the same account or the same server was compromised first, and the file permissions allowed movement between them. Common where several brands sit under one hosting plan for cost reasons.

  4. 04

    An upload path that was never constrained

    A form accepting file uploads that checks the extension in the browser and nothing on the server, writing into a directory the web server will execute from. Usually a feature nobody remembers commissioning.

  5. 05

    A staging copy nobody decommissioned

    The site was compromised at an address like staging or old, running a version from before the last three security releases, sharing the production database because that was convenient at the time.

Afterwards

How to know it is clean rather than quiet

A site stops showing symptoms long before it stops being compromised. The visible payload — the injected links, the redirect, the defacement — is the part the attacker is least attached to; the access they hold is what they came for. Removing the symptom and declaring the incident closed is how the same site is compromised again three weeks later, and the second time it is usually quieter.

Three things establish that a rebuild actually worked. The application came from source control rather than from the compromised filesystem, so nothing unaccounted-for was carried across. Every credential the old environment held has been rotated, including the ones in configuration files, scheduled tasks and third-party integrations. And the entry point has a name — a specific component, a specific account, a specific path — rather than a hypothesis.

Where the entry point genuinely cannot be established — and sometimes it cannot, because the logs had rotated or were never collected — say so in writing rather than closing the incident on an assumption. That record is what justifies the monitoring you put on the rebuilt site, and it is what an assessor or an insurer will ask about later.

Where we help

What we do in an incident

What we do

  • Triage in the first hours, which is when the decisions that cannot be undone get made.
  • Establish the entry point where the evidence supports one, and say plainly when it does not.
  • Specify the rebuild and verify it afterwards, rather than cleaning the compromised environment in place.

Where personal data may have been reached, notification obligations and their timing are set by the data protection regulator. Establish the technical scope quickly, and give it to the people who decide the notification question rather than deciding it inside the technical response.

Primary source

Common questions

What buyers ask

Our host says they cleaned it. Is that enough?+

It removes the symptom and rarely establishes the cause. A hosting provider's cleanup is a scan for known malicious file signatures, which finds the common payloads and misses anything written for your site specifically. Ask them two questions: what the entry point was, and whether they preserved a copy before cleaning. If the answer to the first is "we do not know" and to the second is "no", you have a site that is running again and an incident that is not closed.

How far back should we restore from?+

From before the earliest evidence of change, not from before you noticed. Those dates are frequently weeks apart, because the visible symptom usually appears long after the access was gained. File modification times across the web root give a first indication, and the access logs around that date give a better one. Restoring from yesterday when the entry was six weeks ago restores the compromise along with the content.

Do we have to tell anyone?+

That depends on what was reachable, and it is a question for your legal and compliance function rather than for a technical supplier. What the technical work owes them is a fast, honest scope statement: which systems the compromised access could read, which of them hold personal data, and whether there is evidence of access to it rather than merely the possibility. Give them that early — notification timelines are short and they start from awareness.

Get triage help today

A short triage conversation in the first hours, which is where the decisions that cannot be undone are made.

What happens next

  1. 1We reply within one business day.
  2. 2A 30-minute scoping call with a senior consultant.
  3. 3A written proposal with scope, team, and timeline.

Or reach us directly

By submitting this form you agree to our Terms of Use and Privacy Policy.