> This is part 1 of 2 of the full documentation (pages 1–100 of 192).
> The content is paginated: fetch every part to see all of it.
> Next part: https://docs.stoik.io/llms-full.txt/1
> Page index: https://docs.stoik.io/llms.txt

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
This site has an agentic ask interface you may use to query the documentation dynamically by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.stoik.io/index.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

---

# Help Center

Welcome to Stoïk's help center. Find guides, answers and troubleshooting tips for every Stoïk Protect topic.

<a href="https://docs.stoik.io/onboarding/session" class="button primary">Get started</a><a href="https://docs.stoik.io/cyberattack/contact" class="button secondary" data-icon="triangle-exclamation">Hit by a cyberattack?</a>

<h3 align="center">Explore by topic</h3>

<p align="center">Everything the help center covers, from your first login to incident response.</p>

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-cover data-type="image">Cover image</th><th data-hidden data-type="content-ref"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Cyberattack</td><td>What to do in case of a cyberattack</td><td><a href="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FTssQBVkX43XvKGYVP4Qu%2FCyberattack.png?alt=media&amp;token=5edd56c5-853a-4748-8585-e9f4cf163857">Cyberattack.png</a></td><td><a href="https://docs.stoik.io/cyberattack/contact">CYBERATTACK</a></td><td><a href="https://docs.stoik.io/cyberattack/contact">CYBERATTACK</a></td></tr><tr><td>Onboarding</td><td>How to get started with Stoïk Protect</td><td><a href="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FdQMMuYm76C4PYKyvysy6%2FOnboarding.png?alt=media&amp;token=5bbd6be3-7e22-4c7b-b612-2ae82d5207aa">Onboarding.png</a></td><td><a href="https://docs.stoik.io/onboarding/purpose">Onboarding</a></td><td><a href="https://docs.stoik.io/onboarding/purpose">Onboarding</a></td></tr><tr><td>Prevention tools</td><td>External scan, internal scans, and phishing</td><td><a href="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FmFN6gFGNKjLAFFKwJB4o%2FPrevention%20tools.png?alt=media&amp;token=f93d671f-bd58-4599-997e-3b8fb458c4f4">Prevention tools.png</a></td><td><a href="https://docs.stoik.io/prevention-tools/prevention-tools">Prevention tools</a></td><td><a href="https://docs.stoik.io/prevention-tools/prevention-tools">Prevention tools</a></td></tr><tr><td>Stoïk MDR</td><td>EDR Setup &#x26; Alert Monitoring</td><td><a href="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FjGtJups3hxHbGDA2BIyQ%2FStoik%20MDR.png?alt=media&amp;token=81d15635-34f8-4965-a0c7-aa1761ea1fb2">Stoik MDR.png</a></td><td><a href="https://docs.stoik.io/mdr/deploy">MDR</a></td><td><a href="https://docs.stoik.io/mdr/deploy">MDR</a></td></tr><tr><td>Email Security</td><td>Email Threat &#x26; Compromised Account</td><td><a href="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Ffybr0RD6SAS0WEzpaulB%2FStoik%20Email%20Security.png?alt=media&amp;token=cf9d1b89-f88e-44fe-ae3e-bff906983417">Stoik Email Security.png</a></td><td><a href="https://docs.stoik.io/email-security/microsoft-365">Email Security</a></td><td><a href="https://docs.stoik.io/email-security/microsoft-365">Email Security</a></td></tr><tr><td>Human Security</td><td>Report, training &#x26; phishing simulation</td><td><a href="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F3ugSD5h4BAhCCu0Y61rL%2Fhuman-security.png?alt=media&amp;token=0f337884-9fb9-4a6c-b675-1e409022cd2f">human-security.png</a></td><td><a href="https://docs.stoik.io/human-security/phishing">Human Security</a></td><td><a href="https://docs.stoik.io/human-security/phishing">Human Security</a></td></tr></tbody></table>

***

<h3 align="center">Popular guides</h3>

<p align="center">Quick links to the most common tasks.</p>

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-phone">:phone:</i></h4></td><td><h4>Contact in case of an attack</h4></td><td>Who to call at Stoïk, and when.</td><td><a href="https://docs.stoik.io/cyberattack/contact">Contact in case of an attack</a></td></tr><tr><td><h4><i class="fa-hand-holding-dollar">:hand-holding-dollar:</i></h4></td><td><h4>Compensation</h4></td><td>How your claim is compensated.</td><td><a href="https://docs.stoik.io/cyberattack/compensation">Compensation</a></td></tr><tr><td><h4><i class="fa-calendar-check">:calendar-check:</i></h4></td><td><h4>Onboarding session</h4></td><td>Book your session with our team.</td><td><a href="https://docs.stoik.io/onboarding/session">Onboarding session</a></td></tr><tr><td><h4><i class="fa-shield-halved">:shield-halved:</i></h4></td><td><h4>Deploy the EDR</h4></td><td>Install the MDR agent on your devices.</td><td><a href="https://docs.stoik.io/mdr/deploy">Deploy the EDR</a></td></tr><tr><td><h4><i class="fa-envelope">:envelope:</i></h4></td><td><h4>Email Security setup: Microsoft 365</h4></td><td>Connect your Microsoft 365 tenant.</td><td><a href="https://docs.stoik.io/email-security/microsoft-365">Setup: Microsoft 365</a></td></tr><tr><td><h4><i class="fa-fish">:fish:</i></h4></td><td><h4>Phishing simulation</h4></td><td>Launch your first campaign.</td><td><a href="https://docs.stoik.io/human-security/phishing">Phishing simulation</a></td></tr></tbody></table>


# Contact in case of an attack

In the event of a confirmed or suspected cyberattack, please contact:

* By email at <cert@stoik.io>
* By phone on our 24/7 emergency hotline, at the number listed in your insurance policy. It is also available directly from your Stoïk Protect account:

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FdKptDU9DAbfBaUUvakIC%2Freport-a-claim.png?alt=media&amp;token=5fde1ab9-7ac9-4561-8210-f11c64e3e534" alt=""><figcaption></figcaption></figure>

Stoïk cyber experts will then conduct an investigation to determine if your information system has been compromised.

{% hint style="info" %}
For topics not related to a cyberattack:

* Questions about the Stoïk Protect platform: Contact our cyber consultants at <protect@stoik.io>. You will receive a reply within 24/48 working hours.
* Questions about your insurance policy: Contact your broker.
  {% endhint %}


# Compensation

Helping you to get back on your feet quickly is at the heart of our mission.

That is why our claims management team steps in as soon as the cyberattack has been handled by CERT-Stoïk:

* We send your broker a list of items to send us so that we can calculate the cost of the damage.
* Once we have received these items, we assess the amount of the damage – an independent firm may be appointed to carry out the calculation.
* We send you a letter of acceptance setting out the compensation. Once the letter has been signed, the compensation can be paid.


# Purpose of the account

A dedicated Stoïk Protect account is granted to each Stoïk policyholder: Access is **included at no extra cost** in the insurance policy.

With its four tools (External Scan, Phishing simulation, Active Directory scan and Cloud scan), cyber risks can be identified at a glance, so are areas of the IT system to be tackled as a priority.

The recommendations proposed by the Stoïk Protect platform are intended to enhance your level of security and limit the risk of cyber attacks. Implementing these recommendations is **not contractually mandatory** for the proper functioning of the contract, but an overly compromised risk posture could lead to a reassessment of your insurance contract upon **renewal**.

{% hint style="info" %}
To log into Stoïk Protect: Click on the invitation email received after subscribing, or ask our cyber consultant teams to invite you (contact: <protect@stoik.io>).

\
The login address is [https://app.stoik.io](https://app.stoik.io/)
{% endhint %}

## 1. Intrusion: External Scan and Phishing simulation

This is when the attacker manages to penetrate the information system. This intrusion may exploit:

* A technical vulnerability (open port, exposed service, obsolete software, etc.).
* A human flaw, most often through a social engineering attack such as phishing.

At Stoïk, two tools help reduce the likelihood of intrusion:

1. The External Scan, which identifies technical vulnerabilities visible from the Internet.
2. Phishing simulation campaigns, which raise awareness among your teams and assess their level of resistance to social engineering attempts via email.

## 2. Lateral movement: AD Scan and Cloud Scan

Once inside, the attacker seeks to expand their access and reach more sensitive areas by elevating their privileges. They attempt to map the system and compromise other internal resources.

Stoïk helps reducing an attacker's ability to pivot laterally within the IT system with the:

1. Active Directory Scan, which identifies misconfigurations, accounts with excessive privileges, and security vulnerabilities within the Active Directory.
2. Cloud Scan, which analyses servers and virtual machines hosted in the cloud to detect configuration errors, unsecured access, and vulnerabilities that could expose the sensitive data.

## 3. Impact on objective: CERT Stoïk

The attacker now has access to sensitive data and can carry out their objective: data exfiltration, encryption, sabotage, etc.

At this point, the attack is visible, and the insuree must contact our incident response team (CERT) as soon as possible. CERT team is available 24/7 at <cert@stoik.io> or via the emergency number listed in the insurance contract. This emergency number is also available from the Stoïk Protect account:

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FdKptDU9DAbfBaUUvakIC%2Freport-a-claim.png?alt=media&amp;token=5fde1ab9-7ac9-4561-8210-f11c64e3e534" alt=""><figcaption></figcaption></figure>


# Onboarding session

As a Stoïk policyholder, you can join a training session to help you get started with Stoïk Protect. These sessions will allow you to deploy the prevention tools independently, while also answering all your technical questions.

You can register to a training session from Stoïk Protect's homepage.

{% hint style="info" %}
The information shared during these sessions is technical. It is therefore intended for your IT teams or your IT service provider. For this reason, we recommend that you attend with your technical contact person.
{% endhint %}

The agenda for this training session is as follows:

* Explanation of the emergency procedure in the event of a confirmed or suspected cyberattack.
* Presentation of prevention tools and their deployment.
* Brief explanation of the additional cybersecurity services offered by Stoïk (including MDR).

You can register for an onboarding session from your Stoïk Protect account:

<div align="center"><figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FDdnDXirPgxue2F8lWitC%2Fonboarding-session.png?alt=media&amp;token=cdec8cca-01cc-4d03-ab80-0da982642617" alt=""><figcaption></figcaption></figure></div>


# Reducing the deductible

Installing Stoïk Protect tools entitles you to a 25% discount on your insurance deductible once all the tools are either installed, or marked as irrelevant.

To indicate that a tool is irrelevant for your infrastructure, simply check `This tool is not relevant` in the relevant tool's tab.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FReIMZtCIIFKhK9D91wlb%2Fstoik-protect-ad-scan-not-relevant.png?alt=media&amp;token=f21fcb3e-f146-4175-baa7-95a45aede13c" alt=""><figcaption></figcaption></figure>

Once your tools are categorized as such, you can benefit from a 25% discount on your insurance deductible, and you will no longer receive reminders about their installation.

{% hint style="info" %}
The discount will only be applicable if Stoïk's tools are used.

For instance, if another phishing tool is used but Stoïk's is not, discount won't apply.
{% endhint %}


# Vulnerability notifications

Depending on the severity of the detected vulnerabilities, Stoïk proactively alerts you by email (critical vulnerabilities), or allows you to configure your monthly notifications individually (other vulnerabilities).

{% hint style="info" %}
The information shared is technical. It is therefore intended for your IT teams or your IT service provider, who will correct the vulnerability for you.

To do this, invite them to the platform from the Settings > Users > Invite a user
{% endhint %}

## Prevention tools

* Monthly summary email: An automated email is sent to you on the first of each month, summarizing your activity on Stoïk Protect.
* Critical or High vulnerability notification: If a critical vulnerability is detected on your external network, Stoïk will contact you manually by email, and offer personalized support to resolve it as quickly as possible.
* External Scan report: Users who enabled this option receive an automated email with the External Scan report after each weekly scan, summarizing the vulnerabilities detected on your external surface. It is enabled per user from `Settings` > `Users` (see [User access management](https://docs.stoik.io/onboarding/account/users)).

## Managed services

* Monthly summary email: Identical to the Stoïk Protect email, it is sent to you on the first Monday of each month.
* EDR confirmed alert: If a confirmed threat is detected, Stoïk SOC will send you an email with the alert details and recommended actions.
* Critical Alert: In the event of a major incident requiring immediate action, Stoïk SOC will call you directly.

{% hint style="info" %}
The email address used for manual notifications by Stoik is <alerts@stoik.io>.

The email address used for automatic notifications is <protect@stoik.io>.\
\
**Warning:** any other email address is a potential phishing attempt.
{% endhint %}


# Cockpit

If you are invited to more than one Stoïk Protect account, a **Cockpit** section appears in the sidebar. It gives you a cross-account view of your accounts, without switching spaces one by one.

Typical users: MSPs managing several clients, group CISOs overseeing subsidiaries, IT providers, or holding admins.

The Cockpit includes two views:

* Deployment: deployment status of each cyber tool, for every account you manage
* Alerts: all alerts from all tools and all accounts, in one table

Users with access to a single Stoïk Protect account do not see Cockpit.

## Access Cockpit

1. Sign in to Stoïk Protect with an account that belongs to two or more Protect spaces.
2. In the left sidebar, open the **Cockpit** section
3. Click the **Deployment** or **Alerts** tab

To work inside one client’s space (configure a tool, resolve an alert in context), open that account from the account switcher at the bottom of the sidebar.

## Deployment

Deployment is a matrix: one row per account, one column per tool (for example External Scan, AD scan, Cloud scan, Simulation, MDR, Email Security...).

Each cell shows where that tool stands for that account.

### Tool statuses

| Status                        | Meaning                                                                                                                                |
| ----------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| **Activated**                 | The tool is enabled for this account and is set up and running.                                                                        |
| **Enabled but not activated** | The tool is enabled, but setup is not finished. The client (or you) still needs to complete activation.                                |
| **Not enabled**               | The tool is not enabled for this account.                                                                                              |
| **Not relevant**              | The tool was marked as not applicable for this account (for example already covered elsewhere). This is a valid choice, not a failure. |

### KPIs at the top of the page

Three indicators summarise your fleet:

* **Coverage**: share of possible tool slots that are enabled across all accounts. Use this to see how much of your catalog you deploy.
* **Activation rate**: among enabled tools (excluding Not relevant), how many are activated. Use this to spot tools stuck in setup.
* **Fully protected clients**: share of accounts where every enabled tool is activated. Use this to see how many clients are in a clean state versus needing follow-up.

### Filters and search

Use Tools, Last active, and Added date filters, and the search bar, to narrow the list when you manage many accounts.

## Alerts

The Alerts tab lists alerts from all tools and all accounts you can access in Cockpit.

Each row typically includes the account, the tool, alert details, and an Action column that tells you whether you need to do something.

### Action column

<table><thead><tr><th width="201.0625">Action</th><th>Meaning</th></tr></thead><tbody><tr><td><strong>Informational</strong></td><td>No action is required from you. The tool is fully managed by Stoïk SOC.</td></tr><tr><td><strong>Action required</strong></td><td>You or the client should investigate or remediate.</td></tr></tbody></table>

Use this view to prioritise work across your portfolio instead of opening each account separately.

To handle an alert in full context (history, remediation, settings), open the relevant account and go to the tool’s alerts page.


# Compliance Checklist

The Compliance Checklist is a module in Stoïk Protect that helps you understand which cybersecurity regulations apply to your organisation, and what you need to have in place to meet them. Actions already covered by your Stoïk tools are ticked for you.

## What it covers

* An eligibility check per framework, to determine whether you are in scope and under which category
* A summary of what your category means: supervision, management liability, sanctions and deadlines
* An assessment broken down into concrete actions, mapped to the articles of the regulation
* Automatic ticking of the actions your active Stoïk tools already cover
* A direct link to each tool that would cover a missing action

{% hint style="info" %}
The Compliance Checklist is a self-assessment tool designed to guide you. It is not a legal audit, a certification, or a registration with your national authority.
{% endhint %}

## How it works

{% stepper %}
{% step %}

### Check your eligibility

Confirm a few details about your company. Stoïk pre-fills what it already knows from your insurance contract.
{% endstep %}

{% step %}

### Review your outcome

Stoïk tells you whether the framework applies to you, and what that means in practice.
{% endstep %}

{% step %}

### Complete the assessment

If you are in scope, go through the actions and tick what you already have in place.
{% endstep %}
{% endstepper %}

## Available frameworks

| Framework | Status    |
| --------- | --------- |
| NIS 2     | Available |

## Actions covered by your Stoïk tools

Some actions in an assessment are linked to a Stoïk Protect tool.

* **Tool active**: the action is ticked for you and locked. You see **Stoïk already protects your company**, with a link to the tool.
* **Tool not active**: the action stays open. You see **Stoïk can protect your company**, with a link to set the tool up. You can still tick it yourself if another provider covers it.

All other actions are yours to declare. Stoïk has no visibility over policies, contracts, internal procedures or physical security, so those are always ticked manually.

Your answers are saved automatically as you go. There is no submit button, and you can return to update them at any time.


# NIS 2

The NIS 2 directive requires organisations in critical sectors to put robust cybersecurity measures in place. This page explains what the directive asks for, how to check whether it applies to you, and how to complete the assessment in Stoïk Protect.

## Understanding NIS 2

### What the directive is

NIS 2 ([Directive (EU) 2022/2555](https://eur-lex.europa.eu/eli/dir/2022/2555/oj)) replaces the first NIS directive of 2016. It was adopted in December 2022, and Member States were required to transpose it into national law by **17 October 2024**.

Two things changed significantly compared to NIS 1:

* **The scope widened.** Roughly **18 sectors** are now covered: 11 highly critical sectors in Annex I, and 7 other critical sectors in Annex II. Thousands of mid-sized and large organisations are in scope for the first time, along with public administrations.
* **Management became accountable.** Under [Article 20](https://eur-lex.europa.eu/eli/dir/2022/2555/oj), the management body must approve the cybersecurity measures, oversee their implementation, and can be held personally liable. Executives are also required to follow cybersecurity training.

{% hint style="info" %}
NIS 2 is a **directive**, not a regulation. It sets the objectives; each Member State writes its own law. Your exact deadlines, your registration procedure and any stricter national requirements come from your country, not from the European text.
{% endhint %}

### The three core obligations

<table><thead><tr><th width="242.62109375">Obligation</th><th>What it means</th></tr></thead><tbody><tr><td><strong>Register with your national authority</strong></td><td>Once you have established that you are in scope, you must declare your entity to the competent authority in your country</td></tr><tr><td><strong>Manage your cyber risk</strong></td><td>Put in place the technical, operational and organisational measures of Article 21</td></tr><tr><td><strong>Report significant incidents</strong></td><td>Notify your authority within <strong>24 hours</strong> (early warning), <strong>72 hours</strong> (incident notification) and <strong>one month</strong> (final report)</td></tr></tbody></table>

Failure to comply exposes you to administrative fines of up to **€10M or 2% of global annual turnover** for essential entities, and **€7M or 1.4%** for important entities, whichever is higher in each case.

### Article 21(2): the ten measures

Article 21 is the heart of the directive, and the basis for the assessment in Stoïk Protect.

[Article 21(1)](https://eur-lex.europa.eu/eli/dir/2022/2555/oj) sets the principle: measures must be **appropriate and proportionate**, follow an **all-hazards approach**, and be scaled to your size, your exposure and the severity of the incidents you could suffer. A 60-person manufacturer is not expected to build the same programme as a national energy operator.

[Article 21(2)](https://eur-lex.europa.eu/eli/dir/2022/2555/oj) then lists ten minimum measures. The ten areas of the assessment:

<table><thead><tr><th width="77.8984375"></th><th width="282.796875">Measure</th><th>What the directive asks for</th></tr></thead><tbody><tr><td><strong>(a)</strong></td><td>Risk analysis and information system security policies</td><td>A documented view of your risks, and written security policies approved at the right level</td></tr><tr><td><strong>(b)</strong></td><td>Incident handling</td><td>The ability to detect, qualify, respond to and report incidents</td></tr><tr><td><strong>(c)</strong></td><td>Business continuity</td><td>Backup management, disaster recovery and crisis management</td></tr><tr><td><strong>(d)</strong></td><td>Supply chain security</td><td>Security in your relationships with direct suppliers and service providers</td></tr><tr><td><strong>(e)</strong></td><td>Security in acquisition, development and maintenance</td><td>Including vulnerability handling and disclosure</td></tr><tr><td><strong>(f)</strong></td><td>Assessing the effectiveness of your measures</td><td>Policies and procedures to verify that what you put in place actually works</td></tr><tr><td><strong>(g)</strong></td><td>Basic cyber hygiene practices and cybersecurity training</td><td>For all staff, not only the IT team</td></tr><tr><td><strong>(h)</strong></td><td>Cryptography and encryption</td><td>A policy on where and how encryption is used</td></tr><tr><td><strong>(i)</strong></td><td>HR security, access control and asset management</td><td>Onboarding and offboarding, least privilege, an asset inventory</td></tr><tr><td><strong>(j)</strong></td><td>MFA and secured communications</td><td>Multi-factor or continuous authentication, secure voice, video and text, emergency communications</td></tr></tbody></table>

{% hint style="warning" %}
The directive deliberately names **themes, not controls**. It does not tell you which EDR to buy, how often to patch, or what a password policy should contain. That is why an assessment like the one in Stoïk Protect works at the level of concrete actions: it translates each of the ten themes into things you can actually confirm or put in place.
{% endhint %}

### Where more detail exists

Two sources go further than the directive:

* [**Commission Implementing Regulation (EU) 2024/2690**](https://eur-lex.europa.eu/eli/reg_impl/2024/2690/oj) sets detailed technical and methodological requirements for the Article 21 measures, and the criteria for what counts as a significant incident. It applies to specific cross-border digital entities: DNS providers, TLD registries, cloud and data centre providers, CDNs, managed service and managed security service providers, online marketplaces and search engines, social networks and trust service providers. If you are one of those, this regulation is your reference rather than your national baseline.
* **National referentials.** Member States publish their own set of measures. In France, [ANSSI](https://cyber.gouv.fr/reglementation/cybersecurite-systemes-dinformation/directives-nis-nis2-et-dispositif-saiv/directive-nis-2/) publishes the *Référentiel Cyber France* (ReCyF), which breaks the directive down into 20 security objectives with separate requirements for important and essential entities. These referentials are more granular than the ten areas, and they are what an audit will be based on.

## Step 1: Eligibility check

Open the **Compliance** tab and go to the **NIS 2** card. Three fields determine your outcome:

<table><thead><tr><th width="198.828125">Field</th><th>Options</th></tr></thead><tbody><tr><td><strong>Activity sector</strong></td><td>The sectors listed in Annex I and Annex II of <a href="https://eur-lex.europa.eu/eli/dir/2022/2555/oj">the directive</a>, plus <em>None of the above</em></td></tr><tr><td><strong>Headcount</strong></td><td>Less than 50 · Between 50 and 250 · More than 250</td></tr><tr><td><strong>Aggregated turnover</strong></td><td>Less than €10M · Between €10M and €50M · More than €50M</td></tr></tbody></table>

Fields already known from your insurance contract are pre-filled. You can change them at any time with the **Edit** button.

If your sector is **Digital infrastructure** or **Managed ICT services**, you are also asked for your specific activity (DNS provider, TLD registry, domain name registrar, trust service provider, public electronic communications, cloud computing, data centre, CDN, MSP/MSSP). Some of these activities place you in scope regardless of your size.

{% hint style="info" %}
Only your turnover is requested, not your balance-sheet total. The directive applies the size thresholds of [Commission Recommendation 2003/361/EC](https://eur-lex.europa.eu/eli/reco/2003/361/oj), which considers either one, so we ask for the simpler of the two.
{% endhint %}

### Understanding your outcome

|                           | Essential entity                                  | Important entity                                   | Not concerned                     |
| ------------------------- | ------------------------------------------------- | -------------------------------------------------- | --------------------------------- |
| **Supervision**           | Proactive audit by the authority                  | Audit triggered on incident                        | No NIS 2 supervision              |
| **Management**            | Personal liability (suspension from office)       | Personal liability                                 | No NIS 2 management liability     |
| **Max sanctions**         | €10M or 2% of global revenue, whichever is higher | €7M or 1.4% of global revenue, whichever is higher | No NIS 2 administrative sanctions |
| **Incident notification** | 24h / 72h / 1 month                               | 24h / 72h / 1 month                                | No NIS 2 notification duty        |
| **Deadlines**             | Set by your country's transposition               | Set by your country's transposition                | No NIS 2 registration deadline    |

Essential and Important entities can then open the full assessment with **Complete the full assessment**.

{% hint style="warning" %}
The outcome is indicative. It is based on your sector, your size and your digital activity. It does not account for national specifics such as a designation under the CER directive, a previous status as an Operator of Essential Services, or additional sectors added by your Member State. Confirm your final status with your national authority or your legal counsel.
{% endhint %}

## Step 2: The assessment

The assessment turns the ten measures of Article 21(2) into **40 concrete actions**, four per area. Tick what you already have in place.

| Area                                          | Article  |
| --------------------------------------------- | -------- |
| Risk analysis and security policies           | 21(2)(a) |
| Incident handling and notification            | 21(2)(b) |
| Business continuity and crisis management     | 21(2)(c) |
| Supply chain security                         | 21(2)(d) |
| Lifecycle security and vulnerability handling | 21(2)(e) |
| Effectiveness assessment                      | 21(2)(f) |
| Cyber hygiene and training                    | 21(2)(g) |
| Cryptography and encryption                   | 21(2)(h) |
| HR security, access control and assets        | 21(2)(i) |
| MFA and secured communications                | 21(2)(j) |

### Actions covered by your Stoïk tools

These five actions are ticked automatically when the corresponding tool is active:

| Action                                                           | Covered by                           |
| ---------------------------------------------------------------- | ------------------------------------ |
| Run regular vulnerability scans on internal and external systems | External Scan · AD Scan · Cloud Scan |
| Monitor for leaked credentials tied to company accounts          | Credentials                          |
| Run phishing simulation campaigns                                | Simulation                           |
| Roll out regular cybersecurity awareness training for all staff  | Learning                             |
| Set up detection tooling to catch incidents early                | Stoïk MDR                            |

{% hint style="info" icon="eye" %}
**Disclaimer**<br>

This NIS2 tool is provided for general informational and orientation purposes only. It is intended to give brokers and clients a preliminary indication of whether a company may fall within the scope of the NIS2 Directive and its related obligations.<br>

The results generated by this tool are based on the information provided by the user and Stoïk's current understanding of the applicable legislation, which may be subject to change or differing interpretation at national level. The tool does not constitute legal advice and should not be relied upon as a substitute for a qualified legal assessment.<br>

Stoïk makes no representation or warranty, express or implied, as to the accuracy, completeness, or currentness of the information or results provided, and accepts no liability for any decisions made or actions taken based on the use of this tool.
{% endhint %}

## Official resources

### European texts

| Source                                                                                                   | What it covers                                                                                                                                                                                                        |
| -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Directive (EU) 2022/2555](https://eur-lex.europa.eu/eli/dir/2022/2555/oj)                               | The directive itself: scope criteria (Art. 2, 3), management accountability (Art. 20), risk-management measures (Art. 21), reporting obligations (Art. 23), sanctions (Art. 34) and the sector lists (Annex I and II) |
| [Commission Implementing Regulation (EU) 2024/2690](https://eur-lex.europa.eu/eli/reg_impl/2024/2690/oj) | Detailed technical requirements for Article 21, and the criteria for a significant incident, for cross-border digital entities                                                                                        |
| [Commission Recommendation 2003/361/EC](https://eur-lex.europa.eu/eli/reco/2003/361/oj)                  | The definition of micro, small and medium-sized enterprises used to determine your size category                                                                                                                      |
| [NIS 2 Directive: European Commission](https://digital-strategy.ec.europa.eu/en/policies/nis2-directive) | Overview, FAQ and the state of national transpositions                                                                                                                                                                |
| [ENISA: NIS Directive](https://www.enisa.europa.eu/topics/nis-directive)                                 | Guidance and technical implementation material from the EU cybersecurity agency                                                                                                                                       |

### National authorities

<table><thead><tr><th width="163.39453125">Country</th><th>Authority</th></tr></thead><tbody><tr><td>France</td><td><a href="https://cyber.gouv.fr/reglementation/cybersecurite-systemes-dinformation/directives-nis-nis2-et-dispositif-saiv/directive-nis-2/">ANSSI</a></td></tr><tr><td>Germany</td><td><a href="https://www.bsi.bund.de/">BSI</a></td></tr><tr><td>Belgium</td><td><a href="https://ccb.belgium.be/">Centre for Cybersecurity Belgium</a></td></tr><tr><td>Netherlands</td><td><a href="https://www.digitaltrustcenter.nl/">Digital Trust Center</a></td></tr><tr><td>Spain</td><td><a href="https://www.incibe.es/">INCIBE</a></td></tr></tbody></table>

{% hint style="info" %}
Registration with your national authority, and the deadlines that apply to you, are set by your country's transposition. Check your national authority's page for the current status.
{% endhint %}


# DORA

Since 17 January 2025, the **Digital Operational Resilience Act (DORA, Regulation (EU) 2022/2554)** requires EU financial entities to prove they can withstand, respond to and recover from ICT incidents.

DORA doesn't ask you to buy a specific tool. It asks you to have measures in place **and to prove it**. Stoïk Protect helps with both: its tools cover several technical requirements, and they produce the evidence (reports, history, completion rates) that auditors and supervisors expect.

This article shows which DORA requirements each Stoïk Protect tool helps with, and what stays your responsibility.

{% hint style="info" %}
Stoïk Protect has a dedicated **NIS2** assessment in the **Compliance** tab. There is no DORA assessment in the product yet. DORA and NIS2 overlap heavily on technical measures, and for financial entities DORA replaces the equivalent NIS2 requirements (NIS2, Art. 4). So the NIS2 assessment is still a useful starting point for a gap analysis.
{% endhint %}

## Who is concerned by DORA?

DORA applies to most regulated financial entities: banks, payment and e-money institutions, investment firms, asset managers, insurers and reinsurers, crypto-asset service providers, and others.

Some entities are **exempt**. The main exemption is for insurance and reinsurance intermediaries that are microenterprises, small or medium-sized enterprises (Art. 2(3)). Smaller entities in scope can also apply a **simplified ICT risk management framework** (Art. 16).

If you're not sure whether DORA applies to you, ask your compliance officer or your supervisor (ACPR or AMF in France).

## How Stoïk Protect maps to DORA

DORA has five pillars. Stoïk Protect contributes to four of them.

### 1. ICT risk management (Chapter II, Articles 5 to 16)

This pillar asks you to know your assets, find vulnerabilities, protect your systems, detect anomalies and train your people.

| Stoïk Protect tool               | DORA requirement it helps with                                                                                                                                                                                   | Evidence you can show                                                                |
| -------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ |
| **External Scan**                | Inventory of internet-exposed ICT assets (Art. 8(1), 8(4)). Ongoing identification of vulnerabilities and cyber threats (Art. 8(2)). Monitoring of ICT systems (Art. 9(1)).                                      | Dated scan reports and a history of vulnerabilities and their fixes                  |
| **Active Directory Scan**        | Access limited to what each role needs (Art. 9(4)(c)). Strong authentication (Art. 9(4)(d)). Detection of privileged, dormant or misconfigured accounts.                                                         | AD configuration report and a prioritised list of findings with remediation tracking |
| **Cloud Scan** (AWS, Azure, GCP) | Security and resilience of systems, and prevention of unauthorised access and data leaks (Art. 9(2), 9(3)). Misconfiguration detection on your side of the cloud, where you stay fully responsible (Art. 28(1)). | Cloud posture report per environment                                                 |
| **Credentials monitoring**       | Protection against unauthorised access (Art. 9(3)(b)). Detection of compromised accounts before they're used (Art. 10(1)).                                                                                       | List of leaked credentials detected and passwords rotated                            |
| **Email Security**               | Security of data transfer channels (Art. 9(3)(a)). Reduced risk of unauthorised access and fraud (Art. 9(3)(b)).                                                                                                 | Statistics on blocked threats                                                        |
| **MDR**                          | 24/7 detection of anomalous activity on endpoints (Art. 10(1)). Enough resources to monitor anomalies and cyberattacks (Art. 10(3)). Fast containment (Art. 11(2)).                                              | Activity reports, alerts handled, response times                                     |
| **XDR**                          | Several layers of control, alert thresholds and automatic alerting (Art. 10(2)), across VPN, firewall and cloud data.                                                                                            | Detection rules and alert history                                                    |
| **Training**                     | Mandatory ICT security awareness for all staff (Art. 13(6)). Regular training of the management body (Art. 5(4)).                                                                                                | Completion rates and course certificates                                             |
| **Phishing simulation**          | Awareness programmes (Art. 13(6)). Learning from observed threats (Art. 13(1)).                                                                                                                                  | Click, submission and reporting rates per campaign, with trends over time            |

### 2. ICT incident management and reporting (Chapter III, Articles 17 to 23)

| Stoïk service                                                          | DORA requirement it helps with                                                                                                                                                                                   | Evidence you can show                                                                 |
| ---------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- |
| **MDR / XDR**                                                          | Detecting and logging incidents (Art. 17(2), 17(3)). Early detection matters because DORA's reporting deadlines start from when the incident is classified.                                                      | Alert and incident history                                                            |
| **Phishing Report button**                                             | A simple way for employees to report suspicious emails, which feeds your incident detection process (Art. 17(3)).                                                                                                | Reported emails and how they were handled                                             |
| **Stoïk CERT** (24/7 incident response, included with Stoïk insurance) | Response and recovery (Art. 11). Post-incident review and root-cause analysis (Art. 13(2)). Technical input for major-incident reports (Art. 19(4)): initial notification, intermediate report and final report. | Incident report with timeline, indicators of compromise, root cause and actions taken |

### 3. Digital operational resilience testing (Chapter IV, Articles 24 to 27)

DORA requires a testing programme, reviewed at least once a year, that explicitly includes **vulnerability assessments and scans** (Art. 25(1)).

* **External Scan, Active Directory Scan and Cloud Scan** give you recurring vulnerability assessments that feed this programme (Art. 24(6), 25(1)).
* **Phishing simulations** are a form of scenario-based social-engineering testing (Art. 25(1)).

> Threat-led penetration testing (TLPT, Art. 26 and 27) is only mandatory for financial entities designated by their supervisor. Stoïk Protect doesn't provide it.

### 4. ICT third-party risk (Chapter V, Articles 28 to 44)

You stay fully responsible for your DORA obligations when you use a cloud provider (Art. 28(1)). **Cloud Scan** helps you control the part of that risk that depends on how *you* configured your environment.

It does **not** replace the register of information or the mandatory contract clauses with your ICT providers (Art. 28(3), Art. 30).

### 5. Information sharing (Article 45)

This pillar is optional under DORA. Stoïk Protect doesn't cover it directly.

## What stays your responsibility

Stoïk Protect gives you tools and evidence. Your organisation remains accountable for the following:

* **Governance.** Your management body has final responsibility for ICT risk and must approve the framework (Art. 5(2)).
* **Official reporting.** You notify your supervisor of major incidents (Art. 19). The Stoïk CERT helps you prepare the technical content.
* **Backups and business continuity.** Your backup, restoration and business continuity or disaster recovery policies, and tests of them (Art. 11, 12).
* **ICT provider register and contracts** (Art. 28(3), 30).
* **Encryption policy** (Art. 9(4)).
* **TLPT**, if your supervisor has designated you (Art. 26, 27).

## Getting started

1. Open **Stoïk Protect** and check which tools are already active for your organisation.
2. Activate the tools that match your gaps. External Scan, Active Directory Scan, Training and Phishing simulation are the fastest wins for DORA evidence.
3. Complete the **NIS2 assessment** in the **Compliance** tab for a structured view of your technical measures.
4. Export the reports from each tool and add them to your DORA documentation.

{% hint style="warning" %}
This article is general guidance, not legal advice. References are to the EU texts: Regulation (EU) 2022/2554 (DORA) and its delegated regulations (EU) 2024/1774 and (EU) 2025/301. The exact scope of each service depends on your contract. For your specific obligations, ask your compliance officer or your supervisor.
{% endhint %}


# Account & security

Manage how you and your team sign in to Stoïk Protect, and who can access what.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-right-to-bracket">:right-to-bracket:</i></h4></td><td><h4>Sign in with Google or Microsoft</h4></td><td>Use your work account, no Stoïk password needed.</td><td><a href="https://docs.stoik.io/onboarding/account/sso">Sign in with Google or Microsoft</a></td></tr><tr><td><h4><i class="fa-fingerprint">:fingerprint:</i></h4></td><td><h4>Sign in with a passkey</h4></td><td>Sign in with your device biometrics or a security key.</td><td><a href="https://docs.stoik.io/onboarding/account/passkey">Sign in with a passkey</a></td></tr><tr><td><h4><i class="fa-shield-halved">:shield-halved:</i></h4></td><td><h4>Securing your account</h4></td><td>Set up MFA and manage your recovery codes.</td><td><a href="https://docs.stoik.io/onboarding/account/mfa">Securing your account</a></td></tr><tr><td><h4><i class="fa-users">:users:</i></h4></td><td><h4>User access management</h4></td><td>Invite users, set their rights and delete them.</td><td><a href="https://docs.stoik.io/onboarding/account/users">User access management</a></td></tr><tr><td><h4><i class="fa-lock">:lock:</i></h4></td><td><h4>Requiring MFA</h4></td><td>Enforce MFA on specific users.</td><td><a href="https://docs.stoik.io/onboarding/account/require-mfa">Requiring MFA</a></td></tr></tbody></table>


# Sign in with Google or Microsoft

You can sign in to Stoïk Protect using your Google or Microsoft work account, no Stoïk password needed.

## Signing in

On the Stoïk Protect login page, click **Continue with Google** or **Continue with Microsoft**.

You will be redirected to your provider's sign-in screen. After authenticating, you land straight in Stoïk Protect.

The next time you visit, a **Last used** badge marks the method you used most recently.

## Accepting an invitation via SSO

When a Stoïk Protect admin sends you an invitation, you can join without creating a password:

1. Click the link in your invitation email
2. On the sign-up page, click **Continue with Google** or **Continue with Microsoft**
3. Authenticate with your work account
4. You land directly in Stoïk Protect

Your account must use the same email address as the one the invitation was sent to.

{% hint style="info" icon="lightbulb" %}
If your account has MFA enabled (or required by your admin), you will still be prompted for your second factor after signing in via Google or Microsoft. SSO does not bypass MFA.
{% endhint %}


# Sign in with a passkey

A passkey lets you sign in to Stoïk Protect using your device's biometrics (Face ID, Touch ID, or a hardware security key) instead of a password. Passkeys are phishing-resistant and faster than any code-based method.

You can use a passkey in two ways:

* Passwordless sign-in: tap your fingerprint or face and you're in. No password, no code.
* As a second factor: after entering your password, confirm with your passkey instead of a 6-digit code.

## Setting up a passkey

Open the menu in the lower-left corner of the app, then go to `User settings` > `Security`. In the `Passkeys` section, click **Manage**, then **Add a new device**. You may be asked to confirm your MFA again.

Your browser will prompt you to confirm using your device's authentication method (Face ID, Touch ID, Windows Hello, or a hardware key). Once confirmed, the passkey is saved and ready to use.

You can add multiple passkeys, for example, one on your laptop and one on your phone. Give each one a name so you can tell them apart.

## Signing in with a passkey

On the Stoïk Protect login page, click **Continue with a passkey**.

Your browser will ask you to confirm with your device. After confirming, you land straight in Stoïk Protect, no password needed.

If you have previously signed in with another method, a Last used badge shows on the button you used most recently.

## Using a passkey as a second factor

If your account has MFA enabled, your passkey can serve as the second factor after your password. After entering your email and password, choose Passkey when prompted for your second factor and confirm with your device.

## Managing your passkeys

From `User settings` > `Security` > `Passkeys`, you can:

* Rename a passkey
* Remove a passkey you no longer use

If you remove all passkeys, you will need another sign-in method (password, Google, or Microsoft) to access your account.

## Troubleshooting

<details>

<summary>The passkey prompt doesn't appear</summary>

Make sure you're using a supported browser (Chrome, Safari, Firefox, Edge). The Stoïk desktop app may not support passkeys. Use the browser instead.

</details>

<details>

<summary>"No passkey found for this account"</summary>

You may be on a different device or browser than where you registered the passkey. Try your other devices, or sign in with another method.

</details>

<details>

<summary>You lost the device your passkey was stored on</summary>

Sign in with your password, Google, or Microsoft account. Then go to `User settings` > `Security` > `Passkeys` and remove the lost passkey.

</details>

<details>

<summary>You want to add a passkey from a password manager (1Password, Bitwarden…)</summary>

When your browser shows the passkey prompt, choose your password manager as the storage location.

</details>


# Securing your account

Two-factor authentication (MFA) adds a second layer of security on top of your password. Even if your password is compromised, an attacker cannot sign in without your second factor.

## Setting up MFA

Open the menu in the lower-left corner of the app, then go to `User settings` > `Security`.

Under **Multi-factor authentication**, click **Set up MFA**. You will be asked to choose a method.

### Authenticator app (recommended)

An authenticator app generates a 6-digit code on your phone that refreshes every 30 seconds. It works without a mobile signal and is more secure than SMS.

Popular apps: Google Authenticator, Authy, 1Password, Microsoft Authenticator.

1. Click **Authenticator app**
2. Scan the QR code with your app, or tap **Can't scan it? Enter the setup key** to enter it manually
3. Enter the 6-digit code shown in your app to confirm
4. Save your recovery codes somewhere safe (see below)

### Passkey

A passkey uses your device's biometrics (Face ID, Touch ID) or a hardware security key to confirm your identity, no code to type. You can use it as a second factor after your password, or to sign in without a password at all.

See [Sign in with a passkey](https://docs.stoik.io/onboarding/account/passkey) for setup instructions and full details.

### SMS

A 6-digit code is sent to your phone by text message each time you sign in.

1. Click **Other method → SMS authentication**
2. Enter your phone number
3. Enter the 6-digit code received by SMS to confirm

{% hint style="info" %}
This phone number is tied to your MFA method and stays the same even if you update your profile phone number.
{% endhint %}

## Recovery codes

When you set up an authenticator app, you receive **10 one-time recovery codes**. Each code can be used once to sign in if you no longer have access to your authenticator app. They are only shown once. Save them somewhere safe or download them as a file.

If you use a recovery code to sign in, re-enrol a new authenticator app right afterwards.

## Managing your methods

From `User settings` > `Security`, you can:

* See all active MFA methods and when they were added
* Remove a method (requires confirming your password)

{% hint style="info" %}
If your administrator has required MFA on your account, you cannot remove your last active method.
{% endhint %}

## Signing in with MFA

After entering your email and password, you are prompted for your second factor:

* **Authenticator app**: enter the 6-digit code shown in your app
* **Passkey**: confirm with Face ID, Touch ID, or your hardware key
* **SMS**: enter the 6-digit code sent to your phone
* If you have multiple methods, click **Choose another method** to switch between them
* If you no longer have access to any method, click **No longer have access?** to enter a recovery code


# User access management

{% hint style="info" %}
This page describes an older version of user management. See [Users](https://docs.stoik.io/onboarding/users) for the current `Settings` > `Users` experience.
{% endhint %}

## Invite a user

A user can be invited to Stoik Protect from `Settings` > `Users` and by clicking the `Invite a User` button. You must:

* Enter the invitee's first name, last name, email address, and phone number.
* Grant them at least one access right ("Administrator," "Contract," or "Cyber," see below). The user will then receive the following invitation email:

Until the user accepts the invitation, its status will be `Invitation Pending`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FUxtmMTYyTvluh9ua8PWq%2FScreenshot%202026-01-07%20at%2019.21.02.png?alt=media&amp;token=3e29685a-75c4-4e17-9e75-3457a0a9db07" alt=""><figcaption></figcaption></figure>

## Monitor your users' 2FA status

The `Settings > Users` table includes a `2FA` column showing, for each user, whether they have two-factor authentication enabled on their Stoïk Protect account:

* <i class="fa-check">:check:</i> : the user has 2FA enabled.
* <i class="fa-xmark">:xmark:</i> : the user has not enabled 2FA yet.

To get a quick view of who is unprotected, use the `2FA` filter at the top of the table and select `No`. You'll get the full list of users you should follow up with, or directly enforce MFA on (see [Requiring MFA](https://docs.stoik.io/onboarding/account/require-mfa)).

The 2FA status is read in real time from each user's account. If a user enables or disables 2FA from their own `User settings`, the column updates accordingly.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FYruln8QqNMKnbOgVgrzA%2FScreenshot%202026-05-29%20at%2017.42.48.png?alt=media&amp;token=af4560c5-0e52-4e8c-879c-609499d4e98c" alt=""><figcaption></figcaption></figure>

## User rights and notifications

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FNuVqRGkl24LyFVrGj24A%2Fimage.png?alt=media&amp;token=e973acb1-65e0-49b4-be39-82ab0ca782a4" alt=""><figcaption></figcaption></figure>

### Admin

* Manage user rights: This right allows the user to modify (add or remove) the rights of other Stoik Protect users, including their own. In other words, a user without this right cannot modify their own rights or those of other Stoik Protect users.
* Require MFA: This right forces the user to enable multi-factor authentication. The next time they sign in, they are taken to a mandatory MFA setup screen and cannot access Stoïk Protect until MFA is configured. They also cannot disable their own MFA from their user settings as long as the toggle is on.

### Contracts

* View and edit contracts: This right allows the user to access `Contracts` and `Claims`.
* Receive invoices: This right allows the user to receive invoices (payment reminders) by email as soon as they are generated.

### Cybersecurity: Prevention

* View and edit prevention tools: This right allows the user to access the tabs in the `Prevention` section, namely:
  * External Surface, with external scan.
  * Phishing, with phishing simulation.
  * Active Directory, with Active Directory scan.
  * Cloud Services, with Cloud scan.

{% hint style="info" %}
In the event of a high or critical vulnerability identified by the external scan, users with this right will receive email notifications from Stoïk SOC.
{% endhint %}

* View personal data: This right allows the user to view:
  * `Phishing` > `Results`: the results of employees during phishing simulation campaigns.
  * `External Surface` > `Results` > `Data Leaks`: the identifying data of employees whose data may have been leaked.
* Receive the monthly summary: This right allows the user to receive the monthly summary of the results of the prevention tools by email.
* Receive the scan report after each external scan: This option sends the user an automated email (from <protect@stoik.io>) with the External Scan report after every scan. The External Scan runs weekly, on the day and time configured for your organization, and the report summarizes the vulnerabilities detected on your external surface during that run.

### Cyber: Managed services

This section is only available to clients with a Stoïk MDR contract.

* View and edit Stoik MDR tools: This right allows the user to access the tabs in the `Managed Services` section, namely:
  * Endpoint
  * Identity
* Be notified of alerts by email: the user will be copied on all alerts.
* Be notified of critical alerts by phone
  * Users with priority level 1 will be contacted by phone first, etc.
  * The concept of contact during business hours and non-business hours can also be selected for each user.

{% hint style="info" %}
If no user is available during non-business hours, but a critical vulnerability appears, Stoik will do its best to neutralize the threat until one of your employees intervenes or becomes available. Indeed, the EDRs used by Stoik allow the analysts at the Stoik CERT to temporarily contain most of the detected threats.
{% endhint %}

* Receive the monthly summary: The user will receive a monthly summary of detected alerts.

## Delete a user

Only administrator users can delete a user. To do so, simply select the user in question, click on `Actions`, then `Delete user`.

Certain permissions are essential for a Stoik Protect space: if the user you wish to delete is the only one with these permissions, you must first assign these permissions to another user.


# Requiring MFA

Multi-factor authentication (MFA) adds a second layer of security on top of a user's password. By default, every Stoïk Protect user can enable MFA on their own account from `User settings` > `Security`, in the menu in the lower-left corner of the app.

For compliance reasons (ISO 27001, NIS2, cyber-insurance requirements, internal security policies) administrators can go one step further and **enforce MFA on specific users**: an enforced user is required to set up MFA before they can continue using Stoïk Protect.

## Identify users without 2FA

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FYruln8QqNMKnbOgVgrzA%2FScreenshot%202026-05-29%20at%2017.42.48.png?alt=media&amp;token=af4560c5-0e52-4e8c-879c-609499d4e98c" alt=""><figcaption></figcaption></figure>

Before enforcing MFA, you usually want to know who is still unprotected. The `Settings > Users` table has a dedicated `2FA` column:

* <i class="fa-check">:check:</i> : the user has 2FA enabled.
* <i class="fa-xmark">:xmark:</i> : the user has not enabled 2FA yet.

To get a list of users to follow up with:

1. Go to `Settings > Users`.
2. Open the `2FA` filter at the top of the table.
3. Select `No`.

You now see only users without 2FA. From there, you can either reach out to them directly so they enable it themselves, or open each side panel and tick `Require MFA` to enforce it right away.

## Who can enforce MFA

Only users with the `Manage user rights` permission (i.e. administrators) can require MFA for another user. The `Require MFA` toggle is disabled for users without that permission.

## Requiring MFA for a user

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FuMR57Swm6kO6CO60ShOV%2FScreenshot%202026-05-25%20at%2017.18.46.png?alt=media&amp;token=d6fac69f-d7bc-4218-b134-0d950759e795" alt=""><figcaption></figcaption></figure>

1. Go to `Settings > Users` in Stoïk Protect.
2. Click on the user you want to enforce MFA on. Their side panel opens.
3. In the `Admin` section of the permissions form, tick **Require MFA**.
4. The change is saved with the rest of the permissions form.

{% hint style="info" %}
The same flow lets you turn enforcement off later by unticking the box. Disabling enforcement keeps MFA enabled for the user. They simply regain control over it and can disable it themselves from their settings if they want.
{% endhint %}

## What the enforced user sees

The change applies immediately. As soon as the user interacts with Stoïk Protect, they are sent to the mandatory MFA setup screen and stay there until MFA is fully configured.

* If they are currently using Stoïk Protect, their next click, page change, or refresh takes them to the MFA setup screen. They keep their existing login.
* If they are not currently using Stoïk Protect, they land on the MFA setup screen the first time they open it.
* The setup screen replaces every other page in Stoïk Protect until MFA is configured. The only way forward is to complete the setup.

The setup flow is:

1. **Choose a method**: The user selects between **Authenticator app** (recommended), **Passkey** or **SMS**.
2. **Complete setup**: They follow the steps for their chosen method (scan a QR code for the authenticator app, or enter a phone number for SMS).
3. **Confirm**: They enter a 6-digit code to verify the method.
4. **Done**: They are redirected to the page they were originally trying to reach.

## After MFA is enabled

* They sign in with their password and a second factor (authenticator app code, passkey or SMS code) every time.
* The option to remove their last active MFA method is disabled, with a tooltip explaining that MFA is required by their administrator.

## Edge cases

| Situation                                                     | Behaviour                                                                                                                                                                                                                 |
| ------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| The user already has MFA enabled when the toggle is turned on | They keep using Stoïk Protect as usual. MFA stays required as long as the toggle is on.                                                                                                                                   |
| An admin turns the toggle on for their own account            | Allowed. Their next click, page change, or refresh takes them to the MFA setup screen.                                                                                                                                    |
| The user is currently signed in when the toggle is turned on  | They keep their session. Their next interaction takes them to the MFA setup screen; the rest of Stoïk Protect stays out of reach until MFA is configured.                                                                 |
| A new user is invited with the toggle already on              | Right after they create their account, they land on the forced MFA setup screen.                                                                                                                                          |
| The user can no longer access their phone or app              | If they set up an authenticator app, they can use a recovery code to sign in. If they set up a passkey, they can use another registered passkey or device. Otherwise, ask their admin to temporarily turn the toggle off. |

## Common questions

<details>

<summary>What if the user tries to skip the forced MFA setup?</summary>

The setup screen replaces every other page in Stoïk Protect until MFA is configured. The only ways forward are to complete the setup or to ask their admin to turn enforcement off.

</details>

<details>

<summary>Which MFA methods are supported?</summary>

Authenticator app, SMS, and Passkey are all supported. The authenticator app and passkeys are recommended over SMS: more secure, more reliable, and free to run.

</details>

<details>

<summary>How does the user learn that MFA is now required?</summary>

They discover it in-product: the first time they interact with Stoïk Protect after you save the change, they are sent to the forced MFA setup screen.

</details>

<details>

<summary>What happens to their MFA when I turn the toggle off?</summary>

Their MFA stays enabled. Turning the toggle off simply gives them back control: they can now disable MFA themselves from their settings if they want.

</details>


# Users

Decide who can use your company's Stoïk Protect space, what each person can see and do in it, and which emails and calls they receive.

## How it works

Every user has their own page in `Settings` > `Users`, with four tabs:

* `Identity`: name, contact details, title, and two-factor authentication.
* `Access`: what the user can see and do, tool by tool.
* `Notifications`: the emails and phone calls the user receives.
* `Activity`: every change made to the user's access, and who made it.

Access and notifications are set separately. A user can see a tool without receiving its emails, but they need access to a tool to receive its notifications.

## Titles

A title describes who the user is and pre-fills their access and notifications when you invite them. There are four titles:

* For your own employees: `IT Manager` and `Insurance & Contract`.
* For third parties: `MSP` and `IT Provider`.

A title is applied once, when the user is invited. Changing it afterwards does not change the user's access or notifications: adjust them on the `Access` and `Notifications` tabs.

## In this section

* [Inviting a user](https://docs.stoik.io/onboarding/users/inviting-a-user): invite someone and choose their title.
* [Access](https://docs.stoik.io/onboarding/users/access): set access levels, personal data, and critical actions.
* [Notifications](https://docs.stoik.io/onboarding/users/notifications): choose the emails and phone calls each user receives.
* [Activity](https://docs.stoik.io/onboarding/users/activity): see who changed a user's access, and when.
* [Removing a user](https://docs.stoik.io/onboarding/users/removing-a-user): remove someone's access to your space.
* [Requiring MFA](https://docs.stoik.io/onboarding/account/require-mfa): make two-factor authentication mandatory for a user.


# Inviting a user

Give someone access to your company's Stoïk Protect space. Their title pre-fills their access and notifications, which you can adjust right after.

## Prerequisites

* `Users & access` set to `Write`, with `Invite users` turned on.
* Two-factor authentication enabled on your account.

## Invite a user

1. Go to `Settings` > `Users` and click **Invite a user**.
2. Enter the user's first name, last name, email address, and phone number.
3. Choose their title.
4. Send the invitation.

The user receives an invitation email. You are taken to their `Access` tab, where you can adjust their settings.

## Choosing a title

A title does two things:

* **It shows who the person is.** Anyone looking at the users list knows whether this user is one of your employees or a third party, and what their role is.
* **It comes with a preset of settings.** The title fills in the user's access, critical actions, and notifications, so you don't have to set everything by hand.

Pick the title that matches the person:

* `Internal` **IT Manager**: user in charge of IT and cybersecurity.
* `Internal` **Insurance & Contract**: user in charge of the insurance contract, billing and claims.
* `External` **MSP**: the managed service provider that runs your IT.
* `External` **IT Provider**: another external IT company working for you.

Access levels depend on the title:

| Tool                                                    | IT Manager | Insurance & Contract | MSP   | IT Provider |
| ------------------------------------------------------- | ---------- | -------------------- | ----- | ----------- |
| `Users & access`                                        | Write      | Write                | Write | Write       |
| `Contracts`                                             | Read       | Write                | None  | None        |
| `Claims`                                                | Write      | Write                | Read  | Read        |
| `Compliance`                                            | Write      | Write                | Write | None        |
| `External Scan`, `Credentials`, `AD Scan`, `Cloud Scan` | Write      | None                 | Write | Write       |
| `Simulations`, `Learning`, `Report`                     | Write      | None                 | Write | None        |
| `MDR`, `Email Security`                                 | Write      | None                 | Write | Write       |
| `Remove users`                                          | On         | Off                  | On    | Off         |

## Adjusting a user's settings

The preset is only a starting point. After the invitation, you can change any of the user's settings, one user at a time:

* On the `Access` tab: access levels, `Personal data`, and critical actions.
* On the `Notifications` tab: the emails and phone calls the user receives.

Their title stays the same, so you still know who the person is. Two users with the same title can therefore have different settings.

## Changing a title

You can change a user's title on their `Identity` tab, for example if their role changes. This only updates who the person is. Their access, critical actions, and notifications stay as they are: adjust them on the `Access` and `Notifications` tabs if needed.


# Access

The `Access` tab sets what a user can see and do in Stoïk Protect, tool by tool. Tools are grouped as in the Stoïk Protect menu: Account & Business, Technical Risk, Human Risk, and Managed Defense.

## Access levels

Each tool has one of three levels:

* `None`: the tool is greyed out from the menu. The user cannot open it, and cannot receive its notifications.
* `Read`: the user sees everything in the tool (findings, scores, history, statuses) but cannot change anything. They can receive its notifications.
* `Write`: the user can use the tool and change its settings. However, some critical actions must be granted on top of this access level.

## Users & access

`Users & access` controls who can see and manage the other users of your space. It works slightly differently from other tools: every user can open `Settings` > `Users`, whatever their access level.

* `None`: the user sees only themselves in the list.
* `Read`: the user sees every user of your space.
* `Write`: the user sees every user, and can manage them with the critical actions below.

Every user can change their own notifications: in `Settings` > `Users`, they open their own user and go to the `Notifications` tab.

* To change other users' settings, a user needs `Users & access` set to `Write`, with `Manage other users' access` turned on.
* `Invite users` and `Remove users` also require `Users & access` set to `Write`.

## Personal data

`Personal data` lets the user see employees' personal data in the tools they can access. When it is off:

* `Credentials` is limited to `None`.
* `Simulations`, `Learning`, and `Report` are limited to `Read`.

In `Human Risk`, the user still sees company-wide results, but nothing that identifies an employee:

* Employee lists, in `Overview` and in groups, show sample identities instead of your real employees, and the table is locked.
* Employee detail pages, with each employee's simulation emails and reported emails, cannot be opened.
* When choosing who receives a campaign, names and email addresses are replaced, phone numbers are fully masked, and searching or sorting by name is not available.
* The list of reported emails in `Report` shows sample data instead of the real reports, and the table is locked.
* Results that contain employee data cannot be exported.

## Critical actions

Critical actions are the most sensitive actions in each tool. `Write` access alone is not enough to perform them: each one must be granted on top of it. They appear under a tool once it is set to `Write`, and you turn them on one by one.

> A user needs two-factor authentication enabled to perform a critical action. See [Securing your account](https://docs.stoik.io/onboarding/account/mfa) or [Requiring MFA](https://docs.stoik.io/onboarding/account/require-mfa)

### Account & Business

<table><thead><tr><th width="172.015625">Tool</th><th>Critical action</th><th>What it allows</th></tr></thead><tbody><tr><td>Users &#x26; access</td><td><code>Invite users</code></td><td>Invite new users to your space.</td></tr><tr><td>Users &#x26; access</td><td><code>Manage other users' access</code></td><td>Change other users' title, access, critical actions, and notifications.</td></tr><tr><td>Users &#x26; access</td><td><code>Remove users</code></td><td>Remove users from your space.</td></tr><tr><td>Contracts</td><td><code>Edit company information</code></td><td>Change your company's details.</td></tr><tr><td>Contracts</td><td><code>Edit IBAN and payment mandate</code></td><td>Change the bank details used to pay your contract.</td></tr><tr><td>Claims</td><td><code>Manage claim files</code></td><td>Upload and download the files of a claim.</td></tr></tbody></table>

### Technical Risk

<table><thead><tr><th width="172.2890625">Tool</th><th>Critical action</th><th>What it allows</th></tr></thead><tbody><tr><td>External Scan</td><td>Discard an asset or vulnerability</td><td>Request to discard an asset or a vulnerability, accepting the risk.</td></tr><tr><td>External Scan</td><td>Remove an asset</td><td>Remove a domain name or IP address from the scan.</td></tr><tr><td>Credentials</td><td>Reveal a password</td><td>Display an exposed password in clear text.</td></tr><tr><td>Credentials</td><td>Discard an alert</td><td>Dismiss an exposed credentials alert.</td></tr><tr><td>AD Scan</td><td>Remove an Active Directory domain</td><td>Stop scanning an Active Directory domain.</td></tr><tr><td>Cloud Scan</td><td>Remove a cloud provider</td><td>Disconnect a cloud provider from the scan.</td></tr></tbody></table>

### Human Risk

<table><thead><tr><th width="172.24609375">Tool</th><th>Critical action</th><th>What it allows</th></tr></thead><tbody><tr><td><code>Simulations</code></td><td><code>Turn the payment fraud simulation off</code></td><td>Stop payment fraud simulations for the whole company.</td></tr><tr><td><code>Simulations</code></td><td><code>Stop a running campaign</code></td><td>Stop a phishing campaign in progress.</td></tr><tr><td><code>Simulations</code></td><td><code>Create a custom template</code></td><td>Create your own phishing email templates.</td></tr><tr><td><code>Learning</code></td><td><code>Publish custom content</code></td><td>Publish training content you created.</td></tr><tr><td><code>Learning</code></td><td><code>Turn the training module off</code></td><td>Stop training for the whole company.</td></tr><tr><td><code>Report</code></td><td><code>Triage reported emails</code></td><td>Classify the emails reported by employees.</td></tr><tr><td></td><td><code>Disconnect a tenant</code></td><td>Disconnect your Microsoft 365 or Google Workspace tenant. This stops the three tools at once.</td></tr><tr><td></td><td><code>Disable employees</code></td><td>Exclude employees from simulations and training.</td></tr></tbody></table>

### Managed Defense

<table><thead><tr><th width="173.95703125">Tool</th><th>Critical action</th><th>What it allows</th></tr></thead><tbody><tr><td><code>MDR</code></td><td><code>Uninstall a host</code></td><td>Uninstall the agent from a device.</td></tr><tr><td><code>MDR</code></td><td><code>Send a mobile enrollment</code></td><td>Send employees the email to protect their mobile device.</td></tr><tr><td><code>Email Security</code></td><td><code>Search in mailboxes</code></td><td>Search and read emails in employees' mailboxes. See <strong>Email search</strong>.</td></tr><tr><td></td><td><code>Edit SOC guidelines</code></td><td>Change the instructions the Stoïk SOC follows when handling an alert.</td></tr></tbody></table>

## Common questions

<details>

<summary>Why doesn't a user see a tool in their menu?</summary>

The tool is set to `None` on their `Access` tab. Set it to `Read` or `Write`.

</details>

<details>

<summary>Why can't a user perform an action that is turned on for them?</summary>

They have not enabled two-factor authentication on their account. Ask them to enable it, or make it mandatory. See [Securing your account](https://docs.stoik.io/onboarding/account/mfa).

</details>


# Notifications

The `Notifications` tab sets which emails and phone calls a user receives.

## Who can change notifications

A user's notifications can be changed by two people:

* **The user themselves**, for their own notifications. Every user can do this, whatever their rights: they open their own user in `Settings` > `Users` and go to the `Notifications` tab.
* **A user with `Manage other users' access`**, for any other user. They open that user in `Settings` > `Users` and go to the `Notifications` tab.

In both cases, only the notifications of tools set to `Read` or `Write` for that user can be turned on. A tool set to `None` sends them no notification.

### Account & Business

| Tool        | Notification         | What the user receives                                                     |
| ----------- | -------------------- | -------------------------------------------------------------------------- |
| `Contracts` | `Invoices`           | New invoices and payment reminders.                                        |
| `Contracts` | `Contract lifecycle` | Renewal reminders, onboarding, subjectivities, and payment method updates. |

### Technical Risk

| Tool            | Notification    | What the user receives                               |
| --------------- | --------------- | ---------------------------------------------------- |
| `External Scan` | `Scan report`   | The report sent after each external scan.            |
| `External Scan` | `Vulnerability` | An email when for high and critical vulnerabilities. |

### Human Risk

| Tool                                | Notification            | What the user receives                            |
| ----------------------------------- | ----------------------- | ------------------------------------------------- |
| `Simulations`, `Learning`, `Report` | `Disconnected tenant`   | An email when your mail tenant gets disconnected. |
| `Report`                            | `Reported emails recap` | A recap of the emails reported by employees.      |

### Managed Defense

| Tool                    | Notification                     | What the user receives                                     |
| ----------------------- | -------------------------------- | ---------------------------------------------------------- |
| `MDR`, `Email Security` | `Weekly digest`                  | A weekly digest of the managed defense activity.           |
| `MDR`                   | `Email alerts`                   | Endpoint security alerts by email.                         |
| `MDR`                   | `Critical alerts by phone`       | A phone call for critical alerts, see the next section.    |
| `Email Security`        | `Email security alerts by email` | Compromised mailbox and confirmed suspicious email alerts. |
| `Email Security`        | `Service lifecycle`              | Emails about the Email Security free trial and renewal.    |

If nobody subscribes to `Service lifecycle`, the trial-end emails go to the users with `Email Security` set to `Write`.

### Cross-tool

| Notification      | What the user receives                                                                                               |
| ----------------- | -------------------------------------------------------------------------------------------------------------------- |
| `Monthly summary` | A monthly security summary of your company, with one section per tool the user can access.                           |
| `Claim updates`   | Updates on your company's claims. Available with access to `Claims` or to any `Technical Risk` or `Human Risk` tool. |

## Critical alerts by phone

When a critical alert comes in, the Stoïk SOC calls the users who receive `Critical alerts by phone`. For each of them, set:

* `Contact priority`: users with `Priority 1` are called first, then `Priority 2`, then `Priority 3`.
* `Alert hours`: when the user can be called.
* `Phone number`: the number the SOC calls.

> If no user is available during non-business hours when a critical threat appears, Stoïk does its best to neutralize it until one of your employees becomes available. The EDRs used by Stoïk let the Stoïk CERT analysts temporarily contain most detected threats.

## Notifications that must keep a recipient

Some notifications must always reach at least one user:

* `Invoices`
* `Contract lifecycle`
* `Email alerts` for `MDR`, if your company has Stoïk MDR
* `Email security alerts by email`, if your company has Email Security

You cannot remove a user who is the last recipient of one of them: turn it on for another user first. If one of them has no recipient, a banner warns you in `Settings` > `Users`.

## Companies managed by a broker

If your broker receives Stoïk's communications on your behalf, users cannot turn notifications on. The `Notifications` tab shows a message instead. Notifications already turned on can still be turned off.


# Removing a user

Remove someone's access to your company's Stoïk Protect space.

## Prerequisites

* `Users & access` set to `Write`, with the critical action `Remove users` turned on.
* Two-factor authentication enabled on your account.

## Remove a user

1. Go to `Settings` > `Users`.
2. Select the user.
3. Click `Actions`, then `Delete user`.

The removal is recorded in the user's `Activity` tab.

## When removal is blocked

Stoïk Protect blocks the removal when the user is:

* the last user who can manage other users' access, or
* the last recipient of a notification that must keep one. See [Notifications](https://docs.stoik.io/onboarding/users/notifications).

To remove them, first give the access or the notification to another user.


# Activity

The `Activity` tab lists every change made to a user's access, newest first. Use it to see who gave someone access to what, and when.

## What is recorded

* The user's invitation, first sign-in, and removal.
* Every change to their access levels, critical actions, `Personal data`, notifications, and two-factor authentication requirement, with the value before and after.
* A title applied at invitation, shown as a single entry.
* The critical actions the user performed.

## Who made the change

Each entry names its author:

* a user of your company,
* Stoïk,
* your broker,
* an automatic change made by Stoïk Protect, for example when a tool is set up.

The `Identity` tab also shows who invited the user.

## Who can see it

* Every user can see their own activity.
* To see another user's activity, you need `Users & access` set to `Write`.


# Claims

Each claim has its own page in Stoïk Protect, where you follow the incident with Stoïk's CERT: the actions to carry out, and the files you exchange with the CERT.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-list-check">:list-check:</i></h4></td><td><h4>Action plan</h4></td><td>Follow the actions planned by the CERT and update their status.</td><td><a href="https://docs.stoik.io/onboarding/claims/action-plan">Action plan</a></td></tr><tr><td><h4><i class="fa-folder-open">:folder-open:</i></h4></td><td><h4>Files</h4></td><td>Upload, find and download the files of a claim.</td><td><a href="https://docs.stoik.io/onboarding/claims/files">Files</a></td></tr><tr><td><h4><i class="fa-terminal">:terminal:</i></h4></td><td><h4>Command-line upload</h4></td><td>Send large files straight from a server.</td><td><a href="https://docs.stoik.io/onboarding/claims/command-line-upload">Command-line upload</a></td></tr><tr><td><h4><i class="fa-share-nodes">:share-nodes:</i></h4></td><td><h4>External sharing</h4></td><td>Give a provider without a Stoïk account access to the files.</td><td><a href="https://docs.stoik.io/onboarding/claims/external-sharing">External sharing</a></td></tr></tbody></table>

## Claim page tabs

* `Description`: the details of the claim.
* `Action plan`: the actions planned by Stoïk's CERT. See [Action plan](https://docs.stoik.io/onboarding/claims/action-plan).
* `Processing`: the follow-up of the claim, including the `Data sharing` block for files. See [Files](https://docs.stoik.io/onboarding/claims/files).

To report a new attack, see [Contact in case of an attack](https://docs.stoik.io/cyberattack/contact).


# Action plan

Follow the action plan of your claim in Stoïk Protect, and tell Stoïk's CERT where you stand on each action without sending an email.

## Open the action plan

1. In Stoïk Protect, go to `Claims` and open the claim.
2. Open the `Action plan` tab.

Until the CERT has shared the plan, the tab shows: "Your action plan will appear here once Stoïk's CERT has shared it."

## Read the plan

Each row is one action, with:

* `Status`: `To do`, `In progress` or `Done`.
* `Priority`: from P1 (most urgent) to P3.
* `Action`: the name of the action.
* `Description`: what to do, with the details written by the CERT.
* `Type`: `Investigation`, `Remediation`, `Legal` or `Communication`.
* `Owner`: who carries out the action.

Actions are listed in the order set by the CERT. You can search, filter by status, priority, owner or type, and sort any column.

## Update the status of an action

1. Click the status badge of the action.
2. Pick `To do`, `In progress` or `Done`.

The new status shows at once and is sent to the CERT, along with who changed it. If the update fails, the previous status comes back and an error message tells you why.

## When the plan changes

The CERT keeps the plan up to date during the incident. Their changes appear the next time you open the tab. The plan stays available after the claim is closed.

{% hint style="info" %}
The action plan cannot be downloaded as a PDF from this tab yet. The CERT still sends it by email.
{% endhint %}

For any question on an action, write to <cert@stoik.io>.


# Files

Exchange files with Stoïk's CERT from the claim page: logs, exports, reports or any evidence the investigation needs.

## Prerequisites

* `Claims` set to `Write` on your `Access` tab, with the `Manage claim files` critical action turned on. See [Access](https://docs.stoik.io/onboarding/users/access).
* Two-factor authentication enabled on your account. See [Securing your account](https://docs.stoik.io/onboarding/account/mfa).

## Find the files of a claim

1. In Stoïk Protect, go to `Claims` and open the claim.
2. Open the `Processing` tab and go to the `Data sharing` block.

The table lists every file of the claim, with its name, upload date and time, size, and source: `Company` for files sent by your side, `Analyst` for files sent by the CERT. You can search a file by name, filter on the source, and sort any column.

## Download a file

Click the file in the table. The download starts through a short-lived link.

## Upload files

1. In the `Data sharing` block, click **Upload**.
2. Choose to select files from your computer.
3. Pick one or more files.

Large files are sent in parts. If the connection drops, the upload resumes instead of starting over. You can follow the progress of each file, retry a failed one, or cancel all uploads.

For very large files, or files stored on a server, see [Command-line upload](https://docs.stoik.io/onboarding/claims/command-line-upload).

## Share with a provider

To let someone without a Stoïk account send or download files, see [External sharing](https://docs.stoik.io/onboarding/claims/external-sharing).


# Command-line upload

Copy files from a server straight into the claim storage, without going through a browser. Use it for very large files, or many files at once.

## Prerequisites

* The rights to manage claim files. See [Files](https://docs.stoik.io/onboarding/claims/files).
* One of these tools installed on the server: AWS CLI, AWS Tools for PowerShell, rclone or WinSCP.

## Get the command

{% stepper %}
{% step %}

### Open the command line option

In the claim's `Processing` tab, go to the `Data sharing` block, click **Upload**, then **Upload with a command line**.
{% endstep %}

{% step %}

### Choose your tool and system

Pick the tool installed on your server (`AWS CLI`, `AWS Tools for PowerShell`, `rclone` or `WinSCP`) and its system (`macOS / Linux` or `Windows PowerShell`).
{% endstep %}

{% step %}

### Copy and run the command

Copy the command with the copy button, then run it on the server from the folder that holds your files.
{% endstep %}

{% step %}

### Close the window

The files table reloads, and the files you sent appear in it.
{% endstep %}
{% endstepper %}

## Credentials in the command

* The command holds credentials limited to this claim.
* They are valid for one hour. The window shows until when.
* Once they expire, click **Regenerate** to get a new command. Files keep going to the same place.

{% hint style="warning" %}
Anyone holding the command can upload files to the claim until it expires. Do not share it outside the people who run the upload.
{% endhint %}


# External sharing

Give someone without a Stoïk account, such as your IT provider or a data recovery company, access to the files of a claim through a link. The link expires and can be revoked at any time.

## Prerequisites

* The rights to manage claim files. See [Files](https://docs.stoik.io/onboarding/claims/files).

## Create a link

1. In the claim's `Processing` tab, go to the `Data sharing` block and click **Share**.
2. Enter the `Recipient email`.
3. Choose the `Access`:
   * `Download only`: the recipient sees and downloads the files.
   * `Upload only`: the recipient can only send files.
   * `Download and upload`: both.
4. Choose when the link expires, 30 days at most.
5. Click **Create the link**, then **Copy the link** and send it to the recipient.

## Manage shared access

Every link appears in the `Shared access` table of the `Data sharing` block, with its status: `Active`, `Expired` or `Revoked`.

* While a link is active, you can copy it again.
* Click **Revoke** to cut access at once. Only the person who created the link, or Stoïk, can revoke it.
* All links are revoked automatically when the claim is closed.

## What the recipient sees

The recipient opens the link in a browser, with no account needed. Depending on their access, the page shows the claim's files, a zone to drop files, or both, and until when the link is available.

If the link has expired or been revoked, the page says it no longer works. The recipient must ask you for a new one.

{% hint style="warning" %}
Anyone holding the link reaches the files of the claim until it expires or is revoked. Send it only to its recipient.
{% endhint %}


# Getting started

Stoïk Protect includes prevention tools that detect vulnerabilities on your external surface, Active Directory, cloud environments and exposed credentials.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-globe">:globe:</i></h4></td><td><h4>External Scan</h4></td><td>Detect vulnerabilities on your domains and public IP addresses.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan">External Scan</a></td></tr><tr><td><h4><i class="fa-sitemap">:sitemap:</i></h4></td><td><h4>Active Directory scan</h4></td><td>Audit your Active Directory configuration.</td><td><a href="https://docs.stoik.io/prevention-tools/ad-scan">Active Directory scan</a></td></tr><tr><td><h4><i class="fa-cloud">:cloud:</i></h4></td><td><h4>Cloud scan</h4></td><td>Check your Azure, Entra ID, AWS and GCP environments.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan">Cloud scan</a></td></tr><tr><td><h4><i class="fa-key">:key:</i></h4></td><td><h4>Credentials</h4></td><td>Find leaked passwords and exposed company data.</td><td><a href="https://docs.stoik.io/prevention-tools/credentials">Credentials</a></td></tr></tbody></table>

## Prerequisites

Here is a summary table for each cybersecurity prevention tool available on Stoïk Protect:

* The prerequisites for their configuration
* As well as the technical expertise required to monitor them throughout your contract

| Tool                                                                     | Prerequisites                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              | Monitoring                                                                                                                                                                                 |
| ------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| [External Surface](https://docs.stoik.io/prevention-tools/external-scan) | <p>You need to know:</p><ul><li>Domain names</li><li>Public IP addresses</li></ul>                                                                                                                                                                                                                                                                                                                                                                                                                         | A technical lead is required to fix [vulnerabilities](https://docs.stoik.io/prevention-tools/external-scan/results) throughout the year.                                                   |
| [Phishing](https://docs.stoik.io/human-security/phishing)                | You need an administrator account to [synchronize your mailbox](https://docs.stoik.io/human-security/phishing/setup).                                                                                                                                                                                                                                                                                                                                                                                      | The [creation of phishing campaigns](https://docs.stoik.io/human-security/phishing/campaign) and groups, analysis of results, and training can be carried out without technical expertise. |
| [Active Directory Scan](https://docs.stoik.io/prevention-tools/ad-scan)  | <p>You need:</p><ul><li>The ability to run a script in PowerShell</li><li>To be logged into an account</li><li>Domain administrator</li><li>Enrolled in Active Directory</li><li>Connected to the domain controller</li></ul>                                                                                                                                                                                                                                                                              | A technical lead is required to fix [vulnerabilities](https://docs.stoik.io/prevention-tools/ad-scan/results) throughout the year.                                                         |
| [Cloud Scan](https://docs.stoik.io/prevention-tools/cloud-scan)          | <p>You will need:</p><ul><li>Servers or virtual machines hosted in the cloud (GCP, Azure, AWS) or a Microsoft EntraID subscription</li><li>An administrator account to deploy <a href="https://docs.stoik.io/prevention-tools/cloud-scan/aws">AWS</a>, <a href="https://docs.stoik.io/prevention-tools/cloud-scan/azure">Azure</a>, <a href="https://docs.stoik.io/prevention-tools/cloud-scan/entra-id">EntraID</a>, or <a href="https://docs.stoik.io/prevention-tools/cloud-scan/gcp">GCP</a></li></ul> | A technical lead is required to remediate [vulnerabilities](https://docs.stoik.io/prevention-tools/cloud-scan/results) throughout the year.                                                |


# External Scan

The External Scan identifies vulnerabilities in your information system through an automated, weekly analysis of its external surface. It is non-intrusive and only looks at the infrastructure that is accessible to everyone.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-gear">:gear:</i></h4></td><td><h4>Configuration</h4></td><td>Add your domain names and public IP addresses.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan/configuration">Configuration</a></td></tr><tr><td><h4><i class="fa-chart-line">:chart-line:</i></h4></td><td><h4>Reading results</h4></td><td>Understand your score and each detected vulnerability.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan/results">Reading results</a></td></tr><tr><td><h4><i class="fa-network-wired">:network-wired:</i></h4></td><td><h4>IP addresses and traffic</h4></td><td>Recognize the scanner's traffic and allowlist it.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan/ip-addresses">IP addresses and traffic</a></td></tr><tr><td><h4><i class="fa-ban">:ban:</i></h4></td><td><h4>Excluding false positive assets</h4></td><td>Remove an asset that is not part of your infrastructure.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan/false-positive-assets">Excluding false positive assets</a></td></tr><tr><td><h4><i class="fa-bug-slash">:bug-slash:</i></h4></td><td><h4>Excluding false positive vulnerabilities</h4></td><td>Dispute a vulnerability that is wrong or not yours.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan/false-positive-vulnerabilities">Excluding false positive vulnerabilities</a></td></tr><tr><td><h4><i class="fa-circle-question">:circle-question:</i></h4></td><td><h4>Common questions</h4></td><td>Penetration test, missed vulnerabilities, remote desktop and MFA.</td><td><a href="https://docs.stoik.io/prevention-tools/external-scan/common-questions">Common questions</a></td></tr></tbody></table>

## Attacks exploiting vulnerabilities

In this type of cyberattack, a hacker takes advantage of a vulnerability in your computer system to gain access to it or act within it. These vulnerabilities can be:

* Outdated software that has not been updated.
* Poorly configured services.
* Open ports accessible from the Internet.
* Systems not protected by security mechanisms.

Cybercriminals use automated tools or targeted techniques to detect these flaws, then exploit them to take control of the system, steal data, deploy malware (for example ransomware) or disrupt your business.

Four basic recommendations to protect yourself:

1. Address vulnerabilities detected by the External Scan.
2. Regularly update all systems and software.
3. Limit services exposed to the Internet.
4. Apply security best practices from the initial configuration.

## What the scan checks

You can access the External Scan from the `External Scan` tab in your Stoïk Protect account. It automatically and continuously inspects your resources accessible from the Internet. In particular, it identifies:

* Ports accessible from the outside (potential entry points for an attacker).
* Exposed services and technologies.
* Known vulnerabilities (CVEs) that can be exploited if patches are not applied.

It also examines other external information, such as Internet reputation or the presence of data leaks related to the main domain name. It detects vulnerabilities by connecting to the CVE database and runs port scans to check whether certain ports are open. A detected vulnerability is displayed on the "Infrastructure" tab of your Stoïk Protect platform.

## Methodology and impact on your infrastructure

The scan runs continuously, discreetly and non-intrusively. It generates a very low volume of requests, which limits the risk of detection or blocking by your security tools, and it identifies known vulnerabilities without affecting the operation of your systems.

In more technical terms, it uses your domain name to estimate your attack surface: domains, subdomains, associated IP addresses and the various services. These are all the paths through which a malicious hacker might attempt to gain entry.

The scan is built on an aggregation of open-source tools, mainly Shodan, Nucleus, Censys and Sublist3r. It makes a significant number of queries from an AWS IP range located in Ireland, once a week, during the night from Friday to Saturday.

{% hint style="info" %}
Some Stoïk insurees are regularly scanned by organisations not affiliated with Stoïk.

If a scan affects the availability of your IT system, contact CERT Stoïk, our incident response team, on the emergency number indicated on your insurance policy or at <cert@stoik.io>. The CERT team is available 24/7.

Stoïk also recommends using a web application firewall (WAF) or a solution such as Cloudflare to mitigate distributed denial-of-service (DDoS) attacks.
{% endhint %}

## Customers and prospects

The scan runs even before you subscribe to Stoïk insurance, to determine your eligibility. It does not run the same way for a prospect and for a customer: a prospect needs a fast eligibility decision, while a customer needs comprehensive coverage of vulnerabilities.

The "Prospect" scenario therefore removes several resource-intensive steps:

* The "Bruteforce" task is not enabled: data leaks are not tested.
* The "Technology Discovery" task is not enabled: the technologies used by all customers are no longer checked, to improve scan speed.

The minimum throughput is also increased: more queries are performed per second. Prospect analyses are faster and less thorough than those of customers, which are longer but more comprehensive.


# Configuration

Add your domain names and public IP addresses so the External Scan covers your whole perimeter. You can add assets one by one, or import a long list at once.

## Prerequisites

* At least one domain name, or confirmation that you have none.
* At least one public IP address, or confirmation that you have none.

## Configuration status

Some incidents handled by Stoïk CERT could have been avoided if public IP addresses other than those associated with the domains had been provided. The External Scan is therefore listed as `Not configured` until both prerequisites are met.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FizSvN50eni4DdbOrrftp%2FScreenshot%202026-01-07%20at%2019.41.48.png?alt=media&amp;token=dbe8360c-f448-44fa-a29c-07b0e9f54452" alt=""><figcaption></figcaption></figure>

To configure the External Scan, go to `External Scan` > `Settings`.

## Adding assets

### Adding assets one by one

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FZK9EFxzmzTBA4dk9gYRc%2FScreenshot%202026-05-29%20at%2018.04.26.png?alt=media&amp;token=3b2b7886-7111-4a74-8faa-7bb52df3bc60" alt=""><figcaption></figcaption></figure>

On the Domains or IP addresses card, click `Add` and type the domain name or IP address you want to include. The asset is validated and added to your scan perimeter immediately.

Also add the domain names that you do not use or that you purchased as extras. Attackers may target them too, particularly for impersonation or malicious use.

### Importing in bulk

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F1z8XRFFuh2uXPHePDsnO%2FScreenshot%202026-05-29%20at%2018.05.01.png?alt=media&amp;token=61b9e08c-ccc3-4360-b7ee-6a3429682214" alt=""><figcaption></figcaption></figure>

If you have a long list of domains or IP addresses, you can import up to 500 entries in a single batch.

1. Go to `External Scan` > `Settings`, then click `Import in bulk` on the Domains or IP addresses card.
2. Paste your list in the text area, one entry per line. Copy-pasting a column directly from Excel or a CSV works as-is.
3. Check each line, which is validated in real time:
   * <i class="fa-check">:check:</i> : valid entry, will be submitted.
   * <i class="fa-xmark">:xmark:</i> : invalid format, must be removed.
   * The counters under the field show how many entries are valid vs invalid.
4. Click `Import`.

### Reading the import summary

After you submit, a summary screen breaks down the result into three groups:

* **Added**: entries accepted and immediately included in your scan perimeter. They will appear in the next External Scan run.
* **Pending**: entries received but flagged for manual review by the Stoïk SOC. This typically happens for domains with a high internet reputation, where we double-check that you actually own or are entitled to scan them. They are added to the scan perimeter once approved.
* **Rejected**: entries refused. The most common reasons are:
  * already in your perimeter (duplicate);
  * on Stoïk's blacklist (e.g. public services we never scan);
  * invalid format on the server side.

Close the modal once you have reviewed the summary. To import another batch, reopen `Import in bulk`.

## Good to know

* Results remain visible even if only domains or IP addresses are entered. Warning banners prompt you (on Stoïk Protect and in the monthly summary) to complete the setup.
* The External Scan can also be run on IP addresses only.
* If you purchased your cyber insurance policy via a questionnaire, the IP addresses entered in the questionnaire are automatically imported into Stoïk Protect.
* If you don't have an additional domain name or IP address, check that your infrastructure has no domain name or public IP address. This prevents unnecessary reminders.
* You can only add the domains or IP addresses of a company you are acquiring with a written agreement from that company, together with the list of its assets. Contact <protect@stoik.io>.


# IP addresses and traffic

The IP addresses used by the External Scan cannot be disclosed: they are not static and change regularly, on purpose, to avoid automatic detection or blocking by security systems. This keeps the scans continuous and effective.

{% hint style="info" %}
Stoïk uses the AWS Ireland IP address range to perform external scans. For more details, see the [AWS documentation](https://ip-ranges.amazonaws.com/ip-ranges.json), which lists their IP address ranges.
{% endhint %}

## Allowlisting the scanner

Only the full list of Amazon Ireland IP ranges is available: region `eu-west-1` in the [AWS IP list](https://ip-ranges.amazonaws.com/ip-ranges.json). There is no smaller list.

If you block foreign IP addresses, the scan cannot see your infrastructure, and your results will not reflect your real exposure.

## Checking whether traffic came from the scan

If you saw suspicious traffic or received an abuse complaint:

1. Look up the IP address on [ipinfo.io](https://ipinfo.io). If it does not say Amazon and Dublin, IE, it is not the External Scan (for example, an IP in Frankfurt or at OVH is not Stoïk).
2. Check in Stoïk Protect whether a scan was running at that time.
3. If both match, contact <protect@stoik.io> with the IP address and the date and time, including the timezone.


# Reading results

The `External Scan` > `Results` page gives you a score out of 100 and the list of vulnerabilities the scan detected on your assets.

## External scan score

All the information gathered by the scan is aggregated and weighted into a total score out of 100. The score gives an overview of the overall risk level, calculated from all detected vulnerabilities.

A score of 100/100 means the scan found nothing exploitable from the internet. Usual reasons: a small footprint (for example, just a website hosted by a provider), assets missing from your [configuration](https://docs.stoik.io/prevention-tools/external-scan/configuration), or a firewall that blocks all traffic from abroad, including the scanner in Ireland.

## Detected vulnerabilities

Each vulnerability is displayed below the dashboard. You can filter vulnerabilities by criticality level (low, medium, high, or critical) to see which areas need priority attention. Once a vulnerability is fixed, it is automatically removed during the next scan.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F3Vp3WdOp2Ujdl8zkxcnb%2FScreenshot%202025-12-30%20at%2016.50.24.png?alt=media&amp;token=d29a42aa-1e7b-45e0-a33f-8bdc15368d80" alt=""><figcaption></figcaption></figure>

Vulnerabilities show up while the scan is still running.

Currently, it is not possible to obtain a history of addressed vulnerabilities, to export your assets, or to fetch results through an API.

{% hint style="warning" %}
If a vulnerability cannot be resolved immediately, take temporary measures to minimize the risk of exploitation, such as:

* Applying stricter firewall rules
* Implementing network segmentation

You can contact our cybersecurity engineers at <protect@stoik.io> for advice.
{% endhint %}

## Vulnerability details

Click a vulnerability to open a side panel. It includes:

* **Information:** the affected infrastructure component, a brief summary, the impact of the vulnerability, and remediation recommendations.
* **Why is this in your scope?:** how the scan reached the affected domain or IP, step by step, back to an asset declared in your configuration. For example, an IP is in scope because your domain’s SPF record allows it to send emails. Click any domain or IP to open its details.
* **Email exchanges:** when Stoïk SOC has communicated with your organization about a specific vulnerability (notification, follow-up, remediation guidance), the email thread is displayed directly in the panel. This section only appears when related emails exist.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FLJaFaaaYjmZCJhnNGM3x%2FScreenshot%202025-12-30%20at%2016.51.36.png?alt=media&amp;token=ba261157-4e64-40ca-89c6-1a993f5aec8e" alt=""><figcaption></figcaption></figure>

If, despite the provided recommendations, you are unable to resolve the vulnerability, contact <protect@stoik.io>.

## Rerunning the scan

If you have implemented new remediation measures, click `Rerun scan`. A scan that started before your fix can still report the vulnerability, so rerun it once the fix is in place. If a scan is cancelled, the previous results stay on screen until a new scan completes.

## Scan report

Click `Scan report` to download a report in PDF format, which can be shared externally.

{% hint style="info" %}
To fix a finding, step-by-step hardening guides are available in cyber best practices:

* [Resolve RDP service exposure](https://docs.stoik.io/cyber-best-practices/how-to-resolve-rdp-service-exposure-from-the-internet)
* [Configure SPF](https://docs.stoik.io/cyber-best-practices/harden-your-environment/email-authentication/what-is-spf-and-how-do-i-configure-it)
* [Configure DMARC](https://docs.stoik.io/cyber-best-practices/harden-your-environment/email-authentication/what-is-dmarc-and-how-do-i-configure-it)
  {% endhint %}

## Troubleshooting

<details>

<summary>A fixed vulnerability is still reported</summary>

Go through this list:

* **Did the scan start before the fix?** [Rerun the scan](#rerunning-the-scan).
* **Is it the right port?** The vulnerability details show the exact IP or domain and port. Services often run on a non-standard port (for example, remote desktop on port 7777 instead of 3389).
* **Is the website cached?** Some websites (often WordPress with a cache plugin) keep serving an old copy of their pages, which still shows old versions. Clear the cache, then rerun the scan.
* **Is updating the right fix?** Some vulnerabilities are detected from how the server behaves, not from its version number. Updating does not remove them: follow the remediation in the vulnerability details.
* **Is it based on a version number?** Some findings come from the version number a website displays, and websites sometimes display an outdated number. If you run a fixed version, [request an exclusion](https://docs.stoik.io/prevention-tools/external-scan/false-positive-vulnerabilities).
* **Can you reproduce it?** The vulnerability details include reproduction steps your IT team can run.
* **Is it really yours?** On shared hosting, a finding can belong to another website on the same server. In that case, [request an exclusion](https://docs.stoik.io/prevention-tools/external-scan/false-positive-vulnerabilities).

</details>

<details>

<summary>A vulnerability disappeared from the results</summary>

Missing from one scan is not proof that it is fixed. Common reasons:

* The server did not answer during the scan: it was down, slow, or your hosting provider's protection blocked part of the checks.
* It was excluded: the vulnerability details show who excluded it and why (include dismissed vulnerabilities in the status filter).
* The asset left the scope: a domain was removed or its DNS changed.

For the same reasons, two scans of the same domain can give different results. Stoïk also improves detection regularly, so a newer scan can find more.

</details>

<details>

<summary>A domain or IP I don’t recognize is in the results</summary>

Open its details and read *Why is this in your scope?*. If the asset no longer belongs to you, remove the link at its source. For example, delete the IP from your SPF record or the DNS record pointing to it. It leaves the scope after the next scan.

</details>


# Excluding false positive assets

In rare cases, a domain name, subdomain, IP address or service scanned by the External Scan is not part of your infrastructure. You can then remove or exclude it.

{% hint style="info" %}
Not directly managing a domain name or IP address linked to your infrastructure does not mean it should be removed from the scan. If a vulnerability is detected on a service managed by a third-party provider, it can still impact your IT system: contact the provider to resolve it.
{% endhint %}

## Prerequisites

* The `View and edit prevention tools` right, set in your Stoïk Protect settings.

## Why an asset is in your scope

The scan starts from the domains you entered, then automatically adds what is publicly linked to them:

* subdomains of your domains
* domains that redirect to one of your domains
* IP addresses your domains point to
* mail servers listed in your email configuration (SPF record)

If an asset no longer belongs to you, the cleanest fix is at the source: remove the redirect, the DNS record or the SPF entry. The asset drops out of scope after the next scan. Otherwise, exclude it as described below.

## Removing a domain name or IP address

From the `External Scan` > `Settings` page, you can directly access and remove scanned domain names and IP addresses.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FEOb71TPY7vLuqcHr2RrO%2FScreenshot%202025-12-30%20at%2017.35.16.png?alt=media&amp;token=f31b6491-619e-4619-93ce-c352a9bc8005" alt=""><figcaption></figcaption></figure>

## Excluding an asset

From the `External Scan` > `Assets` page, click the relevant resource.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FZq9xTAaSJkOyxFTibJA7%2FScreenshot%202025-12-30%20at%2017.34.27.png?alt=media&amp;token=95f96a0b-4dcd-4847-9c9f-6611c0c40d5b" alt=""><figcaption></figcaption></figure>

Select the reason for the exclusion, then click `Exclude`. We analyze your request within 48 business hours. Once approved, the asset is removed from the results during the next scan.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FmF9m1mUd9wdIII0rdj37%2FScreenshot%202025-12-30%20at%2017.36.39.png?alt=media&amp;token=3468c87a-460d-4389-ad62-558a509b0bf5" alt=""><figcaption></figcaption></figure>

{% hint style="warning" %}
If a scan is in progress, the `Exclude` button is disabled. Wait for the scan to complete before you submit an exclusion request.
{% endhint %}


# Excluding false positive vulnerabilities

If a vulnerability is a false positive or does not belong to your infrastructure, you can dispute it from the `External Scan` > `Results` page. Before you do, check [A fixed vulnerability is still reported](https://docs.stoik.io/prevention-tools/external-scan/results#a-fixed-vulnerability-is-still-reported).

## Disputing a vulnerability

1. Go to `External Scan` > `Results`.
2. Click the name of the vulnerability.
3. Click `Exclude`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F0ipTT8dciw7oDD1DIvn2%2FScreenshot%202025-12-30%20at%2017.40.42.png?alt=media&amp;token=584fc9c9-6204-4803-9c81-273db005dc9a" alt=""><figcaption></figcaption></figure>

Stoïk's cyber engineers review a dispute request within 48 working hours. Once approved, the vulnerability is removed from the results during the next scan.


# Common questions

Answers to the questions most often asked about the External Scan.

<details>

<summary>Is the External Scan a penetration test?</summary>

No. Here is the difference between the two.

### External Scan

* **Objective:** continuously map and monitor the exposure surface of a computer system.
* **Method:** automated and continuous. It looks at you from the internet, like an attacker would, without passwords, and identifies open ports, exposed services, and known vulnerabilities (CVEs) on assets visible from the internet.
* **Impact on the computer system:** non-intrusive; it does not attempt to exploit vulnerabilities. It is also unobtrusive; it does not disrupt existing systems.

### Penetration test

* **Objective:** produce a detailed report with attack scenarios and advanced recommendations.
* **Method:** performed manually (or semi-automated) by cybersecurity experts, generally on an ad-hoc basis (e.g., 1 to 2 times per year). The penetration test simulates real attacks to exploit vulnerabilities and test the resilience of your systems.
* **Impact on the IT system:** deeper, more intrusive.

{% hint style="info" %}
If you would like to perform a penetration test on your IT system, Stoïk can recommend trusted partners. Contact us at <protect@stoik.io>.
{% endhint %}

</details>

<details>

<summary>Why didn't the scan detect a vulnerability found by an audit or used in an incident?</summary>

The External Scan only checks for known vulnerabilities on what it can reach from the internet. It does not:

* see your internal network or anything behind a login
* analyze source code
* detect that a server has already been compromised

Penetration tests go deeper and can find things the External Scan cannot. If a known vulnerability was reachable from the internet and was not detected, contact <protect@stoik.io> with the asset, the CVE and the date.

</details>

<details>

<summary>Does MFA fix an exposed remote desktop portal?</summary>

No. Remote desktop (RDP) and RD Web Access portals are reported because anyone on the internet can reach them. MFA lowers the risk, but the finding stays until the portal is behind a VPN or no longer reachable from the internet. See [Resolve RDP service exposure](https://docs.stoik.io/cyber-best-practices/how-to-resolve-rdp-service-exposure-from-the-internet).

</details>


# Active Directory scan

The Active Directory scan analyzes the configuration of your on-premises Active Directory with the PingCastle tool and assesses its security level.

An Active Directory (AD) is a directory service provided by Microsoft operating systems. It lets the network administrator centralize user and machine authentication, manage permissions and control access to network resources. Attackers who target a company with an AD aim to become the system administrator: they then control all workstations and servers.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-gear">:gear:</i></h4></td><td><h4>Setup</h4></td><td>Download the script and run it on a domain-joined machine.</td><td><a href="https://docs.stoik.io/prevention-tools/ad-scan/setup">Setup</a></td></tr><tr><td><h4><i class="fa-chart-line">:chart-line:</i></h4></td><td><h4>Reading results</h4></td><td>Understand the risk score and each detected vulnerability.</td><td><a href="https://docs.stoik.io/prevention-tools/ad-scan/results">Reading results</a></td></tr><tr><td><h4><i class="fa-triangle-exclamation">:triangle-exclamation:</i></h4></td><td><h4>Setup errors</h4></td><td>Fix the errors the script prints when it fails.</td><td><a href="https://docs.stoik.io/prevention-tools/ad-scan/setup-errors">Setup errors</a></td></tr><tr><td><h4><i class="fa-circle-question">:circle-question:</i></h4></td><td><h4>Common questions</h4></td><td>What the script is allowed to do and what it sends to Stoïk.</td><td><a href="https://docs.stoik.io/prevention-tools/ad-scan/common-questions">Common questions</a></td></tr></tbody></table>

## What the scan identifies

The scan identifies:

* Outdated objects: systems that are too old or obsolete.
* Overly permissive trust relationships.
* Overly permissive user permissions.
* Anomalies, such as weak password policies.

The results highlight potential vulnerabilities, classified into four levels: low, medium, high and critical. If a vulnerability is detected, click each misconfiguration to see the remediation steps.

{% hint style="info" %}
Two types of Active Directory coexist at Microsoft:

* An on-premises Active Directory, analyzed by the Active Directory scan.
* An Azure Cloud Active Directory, also called Entra ID, whose configurations are analyzed by the [Cloud scan](https://docs.stoik.io/prevention-tools/cloud-scan).
  {% endhint %}


# Setup

Download the AD scan script from Stoïk Protect and run it in PowerShell on a machine joined to your domain.

## Prerequisites

* Windows version:
  * Windows 10 or later on a workstation
  * Windows Server 2016 or later on a server (must have recent security patches applied)
* Internet access: outbound HTTPS (port 443), without SSL inspection, to:
  * `api.prod.ad-scan.cyber.stoik.io`
  * `ad-scan-prod-pingcastle-zip.s3.eu-west-1.amazonaws.com`
  * `ad-scan-prod-reports.s3.eu-west-1.amazonaws.com`
* Enrolled in Active Directory and connected to the domain controller
* Logged in to a domain administrator account
* Unsigned script execution allowed:
  * Option A: `Set-ExecutionPolicy Unrestricted -Scope LocalMachine`
  * Option B: Add the binary file to the ignore list in Defender or a similar solution, then run `Set-ExecutionPolicy Bypass -Scope Process` and run the script.
* Script downloaded within the last 24 hours
* Latest version of PowerShell installed, check with: `$PSVersionTable.PSVersion`

## Running the scan

{% stepper %}
{% step %}

### Open AD Scan

Click `AD Scan`.
{% endstep %}

{% step %}

### Download the script

Click `Download script`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-3c3e9d5fd29ffc83ce39abb83fc910617e57b854%2FCapture%20d%E2%80%99%C3%A9cran%202026-06-09%20%C3%A0%2016.03.34.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
If the AD scan script does not download, your security settings may be blocking it. Check if the file appears in the `Downloads` tab of your file explorer. If the problem persists, modify your security settings to allow downloads.
{% endhint %}
{% endstep %}

{% step %}

### Run the script

Run the script in a PowerShell console from an endpoint that meets the prerequisites above.
{% endstep %}
{% endstepper %}

## Re-running the scan

Updating the AD scan is manual: each new run must be initiated manually. The downloaded script is only valid for 24 hours: after this time, you must download it again.

The scan can be rerun as many times as necessary:

* Under normal operation, we recommend running it monthly.
* During infrastructure security hardening phases, it can be run more frequently to verify that misconfigurations have been corrected.

A script can be run on several machines during its 24 hours, and downloading a new one does not cancel the previous one. Each run replaces the previous results for the domain it scanned. New results show up a few seconds after the script ends.

### Automatic scan with Stoïk MDR

For Stoïk MDR customers on CrowdStrike, Stoïk runs the script once a month through CrowdStrike, on a machine that CrowdStrike identifies as a domain controller. The machine and the date cannot be chosen. The script runs with the domain controller's own account, so it sees everything a Domain Admin sees. For SentinelOne and Microsoft Defender customers, the script must be run manually.

`AD Scan` > `Settings` lists the automatic runs and their status. A failed run is not retried before the next month, but you can run the script manually at any time. The most common cause of failure is a domain controller on Windows Server 2016 or older, see ["Could not create SSL/TLS secure channel"](https://docs.stoik.io/prevention-tools/ad-scan/setup-errors#could-not-create-ssl-tls-secure-channel). To stop automatic runs, contact <protect@stoik.io>.

## Deleting leftover files

The script deletes its files when it ends. If files are left behind (for example after a run with `-KeepLog` or `-NoDelete`), run this command in a PowerShell opened as administrator, from the folder where you ran the script:

```powershell
Remove-Item -Recurse -Force -ErrorAction SilentlyContinue "C:\Temp\Stoik", ".\ad_hc_*"
```

## Troubleshooting

If the script fails, see [Setup errors](https://docs.stoik.io/prevention-tools/ad-scan/setup-errors). For what the script sends to Stoïk and why it is not signed, see [Common questions](https://docs.stoik.io/prevention-tools/ad-scan/common-questions).


# Reading results

The `AD Scan` page shows a risk score and the vulnerabilities found in your Active Directory configuration.

## Risk score

The higher the score, the more closely the Active Directory configuration adheres to security best practices. The aggregate score takes the lowest score among the four security points analyzed by the scan:

* [**Expired Objects**](https://docs.stoik.io/cyber-best-practices/vulnerability-remediation-guides/active-directory-scan/stale-objects)**:** everything related to Active Directory objects and their lifecycle: computer and user creation, delegation.
* **Trusts:** also called "trust relationships", these are the links between Active Directories (reminder: one Active Directory can compromise another in the event of misconfigured trusts).
* [**Privileged Accounts**](https://docs.stoik.io/cyber-best-practices/vulnerability-remediation-guides/active-directory-scan/privileged-accounts)**:** analysis of accounts with administrator rights.
* [**Anomalies**](https://docs.stoik.io/cyber-best-practices/vulnerability-remediation-guides/active-directory-scan/anomalies)**:** rules that are not classified under outdated objects, trusts, or privileged accounts. It analyzes certain misconfigurations, for example weak password policies.

The score comes from PingCastle's points, while the Critical, High, Medium and Low labels are Stoïk's own. So fixing a "critical" finding does not always move the score, and a single category with many points keeps the score at 0. The `Indicators` section of the PingCastle report shows which rules weigh the most.

## Several domains or forests

* PingCastle scans the domains it can reach from the machine, including trusted domains. If a domain is missing, run the script from a machine joined to that domain.
* Each domain has its own results and score: switch domains with the selector at the top of the results page.
* Each run replaces the previous results for that domain: nothing is added up. One machine per domain is enough.
* A domain that no longer exists (for example after a trust was removed) stays on screen: contact <protect@stoik.io> to delete it.

## Vulnerabilities

The scan identifies configuration settings that could expose your environment to security risks. Vulnerabilities can be viewed via three tabs:

* **Cyber projects:** categorizes vulnerabilities by theme: access control, authentication, networks, privileged account management, updates.
* **Attack paths:** gives an overview of risky entry points: compromise of a domain machine, a domain account, or the AD itself.
* **List:** lists all vulnerabilities.

The misconfigured rules found and their criticality levels, ranked from low to critical, are displayed in a table on the `AD Scan` page. Click a vulnerability to see technical information such as its name, detection date, impact, and recommendations for remediating it.

{% hint style="info" %}
To fix a finding, see the [AD Scan remediation guide](https://docs.stoik.io/cyber-best-practices/vulnerability-remediation-guides/active-directory-scan/ad-scan-vulnerability-remediation-guide) for step-by-step remediation of each vulnerability.
{% endhint %}

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F9McDV7aCr4DMFhHU5Pq0%2FScreenshot%202025-12-31%20at%2013.11.27.png?alt=media&amp;token=6dd1c933-b39e-4a4c-86ca-46fc28faef26" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FNieNKdKnMpC7d4ALow5b%2FScreenshot%202025-12-31%20at%2013.11.37.png?alt=media&amp;token=51687b78-f76d-4cd3-bef5-9d63e32489f4" alt=""><figcaption></figcaption></figure>

If SIDHistory has been deleted, you can use a [recovery script](https://github.com/spurrw/restore-sidhistory-access). Ensure that the ACL settings are recreated after the restoration.

Findings cannot be dismissed or marked as accepted risk today. The `AD Scan` page shows the latest report per domain: for older reports, contact <protect@stoik.io>. There is no export or API.

## Troubleshooting

<details>

<summary>A fixed finding is still reported</summary>

* **Did the scan run again?** Results only change when the script runs again (once a month for automatic runs).
* **Which account ran it?** An account without Domain Admin rights cannot see some settings, like password settings objects. This creates findings that come and go between runs (typically `A-NoServicePolicy`). Run the script again with a Domain Admin account, in a PowerShell opened as administrator.
* **What does the PingCastle report say?** Download it from the results page. It lists the exact accounts, machines and GPOs behind each finding.
* **Is it a PingCastle limit?** PingCastle only sees what it checks. For example, it does not detect LAPS replacements like Baramundi, and `A-SmartCardPwdRotation` only checks a domain setting, not whether smart cards are used.
* **Another tool says otherwise?** Stoïk runs PingCastle with its own configuration, so results will not match BloodHound or a different PingCastle version.

</details>


# Setup errors

Fix the most common errors printed by the Active Directory scan script when it fails.

## Reading the script output

Most errors come from an endpoint (workstation or server) that does not meet the [prerequisites](https://docs.stoik.io/prevention-tools/ad-scan/setup#prerequisites). The script output says which one:

* The script runs 10 numbered checks and prints `OK` or `NOK` for each, then a `Mandatory issues` summary. The `NOK` line usually says what to fix.
* If the window closed, open the log file `C:\Temp\Stoik\Stoik_ADScan_<date>.log`. It is kept when the run fails.

## Troubleshooting

<details>

<summary>"Script download age: NOK", "Token expired" or "400 Bad Request"</summary>

The script is more than 24 hours old (check 5 shows its age in hours). Download a new one from Stoïk Protect.

</details>

<details>

<summary>"Domain-joined: NOK"</summary>

The machine is not part of the domain (a server in a workgroup, a personal laptop). Use a domain-joined machine.

If you have no on-premise Active Directory at all (Entra ID only), the AD scan does not apply: use the [Cloud scan](https://docs.stoik.io/prevention-tools/cloud-scan).

</details>

<details>

<summary>"Could not create SSL/TLS secure channel"</summary>

The Stoïk API only accepts TLS 1.2 and 1.3 with recent ciphers. Two usual causes:

* **Old Windows:** Windows Server 2012 and many Windows Server 2016 machines do not offer these ciphers. The easiest fix is to run the script from a more recent domain-joined machine. It does not have to be the domain controller.
* **SSL inspection** on your proxy or firewall: see [PingCastle cannot be downloaded or the report cannot be uploaded](#pingcastle-cannot-be-downloaded-or-the-report-cannot-be-uploaded).

On Windows Server 2016, you can also enable the ciphers from a PowerShell opened as administrator, then run the script again:

```powershell
"TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384", "TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256" | ForEach-Object { Enable-TlsCipherSuite -Name $_ }
```

</details>

<details>

<summary>The script closes automatically after execution</summary>

This means that script execution is restricted on the terminal. Here is a three-step solution:

1. Download the script and go to `Downloads`, then type `powershell -ep bypass` in the search bar:

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FeQr2oHMVVuuNE3wGWQmA%2Fadscan-powershell-ep-bypass.png?alt=media&amp;token=acd7b9c4-f9e0-4b13-a1ce-d86c228bdff3" alt=""><figcaption></figcaption></figure>
2. The PowerShell console should open automatically. Once it has opened, type `& '.\` then press Tab until the correct script file is visible:

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F5YiH2kagFiINDZVXSxnR%2FCapture%20d%E2%80%99e%CC%81cran%202026-07-03%20a%CC%80%2014.46.44.png?alt=media&amp;token=f7205bb2-f424-46c9-bec7-1088ab7906d1" alt=""><figcaption></figcaption></figure>
3. Press Enter.

</details>

<details>

<summary>PingCastle cannot be downloaded or the report cannot be uploaded</summary>

The machine must reach the three hosts listed in the [prerequisites](https://docs.stoik.io/prevention-tools/ad-scan/setup#prerequisites) over HTTPS, without SSL inspection. A proxy that inspects HTTPS traffic to the S3 buckets breaks the download link, even when DNS and TCP connectivity look fine. If that is already done:

* **The `PingCastle.exe` download stalls, then fails with 403** (or the config file downloads but not the `.exe`): a web filter blocks executable files. `PingCastle.exe` is about 230 MB.
* **"Upload failed for ad\_hc\_….xml"** (403 or "Unable to connect"): the firewall blocks uploads to the reports host. If only the `.html` upload failed, results show up but the report cannot be downloaded.
* **Downloading in a browser works, the script does not:** the proxy treats PowerShell traffic differently, usually because of SSL inspection.
* **You go through a proxy:** the script uses the Windows proxy settings. To force one, run it with `-Proxy "http://proxy:8080"`.

To test from the same machine, this command must answer `{"status":"ok"}`:

```powershell
Invoke-WebRequest https://api.prod.ad-scan.cyber.stoik.io/v1/public/health
```

</details>

<details>

<summary>"PingCastle has stopped working" or "No report files found"</summary>

PingCastle is an Active Directory audit tool, so antiviruses often flag it (for example as `Riskware.PingCastle`). Other signs: PingCastle stops with an exit code, or a check mentions that the antivirus may have deleted a file.

* Exclude the script, `PingCastle.exe` and the `C:\Temp\Stoik` folder, or pause the EDR during the scan.
* If your EDR is managed by Stoïk MDR, contact <protect@stoik.io>.
* To help Stoïk investigate, run the script again with `-PCDebug -KeepLog -NoDelete` and send the logs to <protect@stoik.io>.

</details>


# Common questions

Answers to the questions most often asked about the script of the Active Directory scan.

<details>

<summary>Why is the script not signed?</summary>

The script is generated for each download and contains a token unique to your company, valid 24 hours. That is also why it cannot be reused or shared. Allow its execution as described in the [prerequisites](https://docs.stoik.io/prevention-tools/ad-scan/setup#prerequisites).

</details>

<details>

<summary>What does the script send to Stoïk?</summary>

The PingCastle report, plus the error and an excerpt of the log if the run fails. The script downloads PingCastle from Stoïk, runs it on the machine, uploads the report over HTTPS to Stoïk's storage in AWS Ireland, then deletes its files.

</details>

<details>

<summary>Can I run the script offline or check the results before sending them?</summary>

Not possible. The scan only works if the report is sent to Stoïk, and a PingCastle run outside the script does not show up in Stoïk Protect. For exceptional cases, contact <protect@stoik.io>.

</details>


# Cloud scan

The Cloud scan monitors and analyzes the part of your infrastructure located in the cloud. It does not include storage solutions such as OneDrive, SharePoint, or Google Drive.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-microsoft">:microsoft:</i></h4></td><td><h4>Setup: Azure Cloud</h4></td><td>Assign the reader roles on your Azure subscription.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan/azure">Setup: Azure Cloud</a></td></tr><tr><td><h4><i class="fa-microsoft">:microsoft:</i></h4></td><td><h4>Setup: Entra ID</h4></td><td>Grant the reader roles for a Microsoft 365 tenant.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan/entra-id">Setup: Entra ID</a></td></tr><tr><td><h4><i class="fa-aws">:aws:</i></h4></td><td><h4>Setup: AWS</h4></td><td>Create the read-only policy and role for your AWS account.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan/aws">Setup: AWS</a></td></tr><tr><td><h4><i class="fa-google">:google:</i></h4></td><td><h4>Setup: GCP</h4></td><td>Create the service account for your Google Cloud project.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan/gcp">Setup: GCP</a></td></tr><tr><td><h4><i class="fa-chart-line">:chart-line:</i></h4></td><td><h4>Reading results</h4></td><td>Understand your score and each detected misconfiguration.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan/results">Reading results</a></td></tr><tr><td><h4><i class="fa-user-shield">:user-shield:</i></h4></td><td><h4>MFA status</h4></td><td>Check which employees have enabled MFA in Azure.</td><td><a href="https://docs.stoik.io/prevention-tools/cloud-scan/azure/mfa-status">MFA status</a></td></tr></tbody></table>

## How it works

The Cloud scan uses the open-source tool ScoutSuite, and is available for the leading cloud providers: Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP). It analyzes elements of your cloud configuration deemed critical from a cybersecurity perspective:

* **User access methods and rights:** verify that user rights and authentication methods are properly hardened.
* **Password policy:** audit the password policy and ensure it is not overly permissive.
* **Database exposure:** validate that databases are not publicly exposed on the internet and analyze access rights.
* **Firewall rules:** audit access rules to the cloud infrastructure.
* **Backup policy:** verify that backups are performed regularly.

{% hint style="info" %}
Stoïk Protect only has read access to the cloud infrastructure configuration, enabling technical auditing. It does not access data stored in the cloud (company data, emails) and cannot modify the infrastructure.
{% endhint %}

## Scan frequency

The scan runs about every 24 hours, counted from when you finished the setup, without any action required from you. To refresh the results right away, restart the scan from `Cloud Scan` > `Settings`.

## Entra ID

The Cloud scan analyzes Active Directories located in the cloud (also known as Entra IDs). When only Office 365 via Entra ID is scanned, the analysis focuses on aspects related to your mailbox, such as detecting users who have not enabled two-factor authentication.

If you only have Microsoft 365, with no Azure servers, use the [Entra ID](https://docs.stoik.io/prevention-tools/cloud-scan/entra-id) provider. In `Cloud Scan` > `Settings`, Entra ID-only providers are displayed under the Azure logo.

## Several accounts, projects or subscriptions

Each AWS account, GCP project or Azure subscription is set up separately, from `Cloud Scan` > `Settings` > `New Provider`. There is no organization-wide setup and no API to automate it.

A tenant or cloud account can only belong to one Stoïk Protect company. If you get "This tenant / account is already linked to another company", or if you added the wrong tenant or the same tenant twice, contact <protect@stoik.io> with the tenant or account ID. Do not add the same tenant again on top.

## What is not covered

* Google Workspace (Gmail, Drive, Calendar): the Cloud scan covers Google Cloud Platform only.
* Azure DevOps, and SaaS tools hosted by a vendor.
* Entra ID app registrations, service principals and OAuth apps.
* A full review of Conditional Access policies: they are only read for the MFA findings.
* Other clouds, such as Scaleway, OVH, Orange or Magenta Cloud.


# Setup: Azure Cloud

Give the Cloud scan read access to an Azure subscription by assigning reader roles to the Stoïk Cloud Scanner application.

## Prerequisites

* Servers and/or virtual machines hosted in Azure, and an administrator account.
* A technical advisor is required for initial configuration and ongoing supervision.
* One person does the whole setup, from start to finish: a Stoïk Protect user allowed to edit prevention tools, who is also an administrator of the Azure account. Most Azure failures come from two people splitting the flow.
* Signed in to a single Microsoft tenant: the one to scan. If the setup is stuck, restart it in a private browser window, signed in to that tenant only.

If you only have Office 365 or Entra ID licenses, see [Setup: Entra ID](https://docs.stoik.io/prevention-tools/cloud-scan/entra-id) instead.

## Setup

{% stepper %}
{% step %}

### Select Microsoft Azure

* If you haven't set a provider yet: go to `Cloud Scan` > `Launch Cloud Scan` > `Azure`.
* If you already have a provider: go to `Cloud Scan` > `Settings` > `New Provider` > `Azure`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FoQwu5YkYywfunHxjbihU%2Fstoik-protect-cloudscan-azure-setup.png?alt=media&amp;token=4f7d918d-4749-4216-b057-74fa267a6606" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Create new roles

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FOu6QiRSWQUTtmv8nipXf%2Fstoik-protect-cloudscan-azure-setup-step-one.png?alt=media&amp;token=3ee7937d-08ad-40da-b2c3-6b85e63075ec" alt=""><figcaption></figcaption></figure>

1. Go to the [Microsoft Azure portal](https://portal.azure.com/auth/login/) and log in with your administrator account.
2. Search for `Subscriptions` in the search bar, then click on the subscription to scan.

   <div align="left"><figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FGXhzLVcnk3NXmOSpBhc3%2Fazure-search-subscriptions.png?alt=media&amp;token=3811f8d2-b23e-4334-9d04-cf7b96c9552f" alt="" width="563"><figcaption></figcaption></figure></div>

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FsdXPrOYAmq34SA053yB9%2Fazure-subscriptions.png?alt=media&amp;token=2cd4a6f4-ed6a-48a9-9c5d-1f7e69ab86ff" alt=""><figcaption></figcaption></figure>
3. Click `Access control (IAM)` in the left menu, then click `Add`, then `Add role assignment`.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FJWr92yL06LGvlAD7qpRS%2Fazure-subscriptions-add-role-assignment.png?alt=media&amp;token=3fd7955e-c481-4add-b739-c7c655e83d3a" alt=""><figcaption></figcaption></figure>
4. Search for the `Reader` role, select it, then go to the `Members` tab.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FMSN4E1BtIFsKsuAUEAy8%2Fazure-add-role-assignment.png?alt=media&amp;token=f7e63c9e-3f7b-4054-9c9f-91d96733c3cc" alt=""><figcaption></figcaption></figure>
5. Click `Select members`, search for `Stoïk Cloud Scanner`, select it, then click `Select`.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fy4LLrmvERZiGT0x4kLQc%2Fazure-select-members-role-assignment.png?alt=media&amp;token=25861f02-defa-4f05-bfef-855ad3824b0c" alt=""><figcaption></figcaption></figure>
6. Click `Review + assign` twice to confirm.
7. Repeat steps 4 to 6 for the `Security Reader` and `Log Analytics Reader` roles.
   {% endstep %}

{% step %}

### Grant permissions

<div align="left"><figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-f651e9e32a49814f6f3ca25e5bc9d9d065eea011%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-27%20%C3%A0%2017.04.44.png?alt=media" alt="" width="563"><figcaption></figcaption></figure></div>

1. In the left-hand menu, go to `Microsoft Entra ID` > `Roles and administrators`.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-04fc946096f95da5162151c5a51e74341bbb8967%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.23.34.png?alt=media" alt=""><figcaption></figcaption></figure>
2. Search for the `Global Reader` role and click on it.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-a92451301f63e3b3c1d93cec35588d3169146816%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.24.23.png?alt=media" alt=""><figcaption></figcaption></figure>
3. Click `Add assignments`, search for `Stoïk Cloud Scanner`, select it, then click `Add`.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-92d583a4bfc037d195a2c22e8a02fcff5b22f205%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.25.49.png?alt=media" alt=""><figcaption></figcaption></figure>

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-9cdf09a562407940b88cde427056f5d424b7b2b3%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.28.49.png?alt=media" alt=""><figcaption></figcaption></figure>
4. Repeat steps 2 and 3 for the `Security Reader` role.

{% hint style="warning" %}
Do not grant additional rights such as `Global Secure Access Administrator`, otherwise it will not work.
{% endhint %}
{% endstep %}

{% step %}

### Complete setup

Click `Complete setup` in the Stoïk Protect console.

<div align="left"><figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-f651e9e32a49814f6f3ca25e5bc9d9d065eea011%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-27%20%C3%A0%2017.04.44.png?alt=media" alt="" width="563"><figcaption></figcaption></figure></div>

{% hint style="info" %}
Refresh the Stoïk Protect page if necessary.
{% endhint %}
{% endstep %}
{% endstepper %}

## Several subscriptions

To scan several subscriptions, assign the three subscription roles (`Reader`, `Security Reader`, `Log Analytics Reader`) on each of them.

## Troubleshooting

<details>

<summary>The setup failed, then worked when I retried</summary>

New roles and permissions take a few minutes to apply on Azure. Wait 5 minutes and click again.

</details>

<details>

<summary>"Some roles are not configured correctly", or the setup goes back to the start</summary>

The check expects exactly the roles above: a missing role fails, and so does an extra one. Common mistakes:

* Permissions granted in Entra ID but roles never assigned.
* `Reports Reader` added as a role: remove it.
* `Entra Backup Reader` or a Global Secure Access role instead of `Security Reader`.
* `Monitoring Reader` instead of `Log Analytics Reader`.
* Roles on one subscription but not on the others.

Error messages use the English role names, even when your portal is in another language.

</details>

<details>

<summary>"Stoïk Cloud Scanner" can't be found</summary>

See ["Bad Tenant" error](https://docs.stoik.io/prevention-tools/cloud-scan/azure/bad-tenant).

</details>

<details>

<summary>Policy.Read.All cannot be granted, or a permission is missing</summary>

See [Policy.Read.All cannot be granted](https://docs.stoik.io/pages/o078MxFZGrJa3CHRphbh#policy.read.all-cannot-be-granted-or-a-permission-is-missing).

</details>


# "Bad Tenant" error

The `BAD TENANT` error appears when the Azure setup of the Cloud scan runs on a different tenant than the one you authorized.

## Cause

When running the automatic script for the Azure Cloud scan, the following error may appear:

`BAD TENANT - Stoïk Cloud Scanner application not found, please see this link for more information`

This error occurs when the **Stoïk application has not been registered correctly on your Azure environment**. The main reason is that a different environment was selected during the first stage of setup and when the script was run.

## Example

The example below illustrates a problematic setup:

{% stepper %}
{% step %}

### Select the provider

The user selects the provider `Microsoft Azure`.

<figure><img src="https://content.gitbook.com/content/aDGGQ49Oui29Ft0kJjSm/blobs/NWPK4gXD1wKs8U1vuOPE/image%20png%20Oct%2003%202024%2009%2016%2002%204325%20AM.png" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Choose the account

The user selects their `xavier@bar.test` account associated with the tenant `bar.test`.

<figure><img src="https://content.gitbook.com/content/aDGGQ49Oui29Ft0kJjSm/blobs/IJMiaLxdwfQvqyMiueVz/image%20png%20Oct%2003%202024%2009%2014%2025%204762%20AM.png" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Request the application registration

The Stoïk Cloud Scanner application asks to register on the tenant `bar.test`.

<figure><img src="https://content.gitbook.com/content/aDGGQ49Oui29Ft0kJjSm/blobs/V2k0lLoguK5BYclSqEXW/image%20png%20Oct%2003%202024%2009%2021%2042%203314%20AM.png" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Run the script on a different tenant

The user tries to run the script on the tenant `foo.test` (shown at the top right of the image). Because the Stoïk Cloud Scanner application was authorized on a different tenant (`bar.test`), the script cannot assign the necessary roles on `foo.test`.

<figure><img src="https://content.gitbook.com/content/aDGGQ49Oui29Ft0kJjSm/blobs/9xq9gyyZ3EhqYwuXS0B8/image%20png%20Oct%2003%202024%2009%2034%2044%202548%20AM.png" alt=""><figcaption></figcaption></figure>
{% endstep %}
{% endstepper %}

## Solution

* Run the script on the same Azure tenant that was used when authorizing the Stoïk application (the tenant selected in step 2 above).
* If you are unsure which tenant was used:
  * Disconnect from all your accounts on <https://portal.azure.com>.
  * Then restart the installation of the tool on Stoïk Protect.

{% hint style="info" %}
If the issue persists, contact Stoïk at <protect@stoik.io> for further assistance.
{% endhint %}


# MFA status

Check which employees have enabled MFA in Microsoft Azure, from the Entra ID portal.

## Prerequisites

* An administrator account on the [Microsoft Azure portal](https://portal.azure.com).

## Checking MFA status

{% stepper %}
{% step %}

### Log in to Azure

Go to the [Microsoft Azure portal](https://portal.azure.com) and log in with your administrator account.
{% endstep %}

{% step %}

### Open Microsoft Entra ID

In the left-hand menu, click `Microsoft Entra ID`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FlgnuoBc4blVhx1KBnc8k%2Fazure-entra-id.png?alt=media&amp;token=68972a0c-c9a5-4c7d-8d55-13b0b54e02e6" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Open Users

Click `Manage` > `Users`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fp2u712uTGfJOB3AopLm7%2Fentra-id-manage-users.png?alt=media&amp;token=9cfbcb75-b84a-496e-ae3b-049bbf32077e" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Open MFA by user

Click on `Per-user MFA`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FJPtwlPwQOMve5lfMejBl%2Fentra-per-user-mfa.png?alt=media&amp;token=25cb5971-3a62-43d3-ac0b-0f8735ee010b" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Read the results

You will see the MFA status list.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FZ5BKfOo6jO5KkVhrE5VE%2Fentra-mfa.png?alt=media&amp;token=acf4c4db-0af8-4aa5-9d02-dded0f788591" alt=""><figcaption></figcaption></figure>
{% endstep %}
{% endstepper %}


# Setup: Entra ID

Give the Cloud scan read access to your Microsoft 365 tenant by assigning two reader roles to the Stoïk Cloud Scanner application.

## Prerequisites

* An Entra ID (Active Directory in the cloud) account with administrator access.
* A technical advisor is required for ongoing supervision and vulnerability remediation.
* One person does the whole setup, from start to finish: a Stoïk Protect user allowed to edit prevention tools, who is also an administrator of the Entra ID tenant.
* Signed in to a single Microsoft tenant: the one to scan. If the setup is stuck, restart it in a private browser window, signed in to that tenant only.

## Setup

{% stepper %}
{% step %}

### Select Entra ID

* If you haven't set a provider yet: go to `Cloud Scan` > `Launch Cloud Scan` > `Entra ID only`.
* If you already have a provider: go to `Cloud Scan` > `Settings` > `New Provider` > `Entra ID only`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-48364340474f659a77ae34970c101269371e0e5d%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.12.37.png?alt=media" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Grant permissions

1. In the left-hand menu, go to `Microsoft Entra ID` > `Roles and administrators`.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-04fc946096f95da5162151c5a51e74341bbb8967%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.23.34.png?alt=media" alt=""><figcaption></figcaption></figure>
2. Search for the `Global Reader` role and click on it.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-a92451301f63e3b3c1d93cec35588d3169146816%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.24.23.png?alt=media" alt=""><figcaption></figcaption></figure>
3. Click `Add assignments`, search for `Stoïk Cloud Scanner`, select it, then click `Add`.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-92d583a4bfc037d195a2c22e8a02fcff5b22f205%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.25.49.png?alt=media" alt=""><figcaption></figcaption></figure>

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-9cdf09a562407940b88cde427056f5d424b7b2b3%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.28.49.png?alt=media" alt=""><figcaption></figcaption></figure>
4. Repeat steps 2 and 3 for the `Security Reader` role.

{% hint style="warning" %}
Do not grant additional rights such as `Global Secure Access Administrator`, otherwise it will not work.
{% endhint %}
{% endstep %}

{% step %}

### Complete setup

Click `Finish` in the Stoïk Protect console.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-5438feba60318398574119da452e9fa837bbf50d%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2017.30.42.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Refresh the Stoïk Protect page if necessary.
{% endhint %}
{% endstep %}
{% endstepper %}

## Troubleshooting

<details>

<summary>Policy.Read.All cannot be granted, or a permission is missing</summary>

If you get "Policy.Read.All cannot be granted", or if `Reports.Read.All` or `UserAuthenticationMethod.Read.All` is missing, these permissions cannot be added by hand. Grant admin consent to the Stoïk application again:

1. In [portal.azure.com](https://portal.azure.com), search for **Stoïk Cloud Scanner** (`Enterprise applications`).
2. Open `Security` > `Permissions`.
3. Click **Grant admin consent for \<organization>** and accept the popup.

Consent must cover the whole organization, not only the admin's own account. Without `Policy.Read.All`, Stoïk cannot read your Conditional Access policies, which affects the [MFA findings](https://docs.stoik.io/prevention-tools/cloud-scan/results#mfa-findings-entra-id).

</details>


# Setup: AWS

Give the Cloud scan read-only access to an AWS account by creating a policy and a role, then pasting the role ARN in Stoïk Protect.

## Prerequisites

* An administrator account on the AWS account to scan.
* If you use AWS Organizations, your Service Control Policies (SCPs) must not block the read-only permissions below. SCPs are the most common cause of setup failures, see [Troubleshooting](#troubleshooting).
* A technical advisor is required for ongoing supervision and vulnerability remediation.
* One person does the whole setup, from start to finish: a Stoïk Protect user allowed to edit prevention tools, who is also an administrator of the AWS account.

## Setup

{% stepper %}
{% step %}

### Select Amazon Web Services

Go to `Cloud Scan` > `Settings`, click `New Provider`, and select `AWS`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FL48UGpzqbZCjyJWwIZd1%2FScreenshot%202026-04-07%20at%2018.48.29.png?alt=media&amp;token=205ab9b0-31ea-4328-8a16-2afa1bdc2162" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Create a new policy

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FivaUKKzBa0gW0H6BRi2M%2FScreenshot%202026-04-07%20at%2018.48.55.png?alt=media&amp;token=a8b07550-2f4f-439f-a88b-17314ceeac29" alt=""><figcaption></figcaption></figure>

#### Configure the policy

1. Go to the [AWS policy creator](https://console.aws.amazon.com/iam/home?#/policies$new?step=edit) and log in with your administrator account.
2. Click the `JSON` tab and paste the following policy into the editor:

   ```json
   {
     "Version": "2012-10-17",
     "Statement": [
       {
         "Effect": "Allow",
         "Action": [
           "ecr:BatchGetRepositoryScanningConfiguration",
           "ecr:DescribeImageScanFindings",
           "ecr:DescribeImages",
           "dynamodb:ListTagsOfResource",
           "ecr:ListTagsForResource",
           "inspector2:BatchGet*",
           "inspector2:Get*",
           "inspector2:Describe*",
           "inspector2:List*",
           "sqs:ListQueueTags",
           "sso:Describe*",
           "sso:List*"
         ],
         "Resource": "*"
       },
       {
         "Effect": "Deny",
         "Action": [
           "datapipeline:EvaluateExpression",
           "datapipeline:QueryObjects",
           "rds:DownloadDBLogFilePortion"
         ],
         "Resource": "*"
       }
     ]
   }
   ```
3. Click `Next`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fx9ykOc6VoV6cuAGIaXYb%2Faws-create-policy.png?alt=media&amp;token=140fe2ed-5cbd-43b0-b296-6efb60a60018" alt=""><figcaption></figcaption></figure>

#### Save the policy

1. Name the policy `StoikAdditionalPermissions`.
2. Click `Create policy`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FNpe6hR0vOQiCv87lK1DO%2Faws-create-policy-validation.png?alt=media&amp;token=855b079e-873d-4582-a215-e170c64249c7" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FoANQIZUox1pOS263nogA%2Faws-create-policy-button.png?alt=media&amp;token=2d52294d-7aeb-4ee8-92cb-fcd24c57a328" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Create a new role

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FfUh2nx1vfzxqe3rKTvTZ%2FScreenshot%202026-04-07%20at%2018.50.39.png?alt=media&amp;token=0307ddd7-b07d-4d5e-9c08-addbf5814ebc" alt=""><figcaption></figcaption></figure>

#### Configure the role trust

1. Go to the [AWS role creator](https://us-east-1.console.aws.amazon.com/iamv2/home#/roles/create?step=selectEntities).
2. Click `AWS account`, then `Another AWS account`, and enter the following account ID:

   ```
   436578779424
   ```
3. Select `Require external ID` and copy the value displayed in the Stoïk Protect console.
4. Click `Next`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FPtqCLbhc6EubRg0FAKJh%2Faws-create-role.png?alt=media&amp;token=ec01ef82-a562-4c76-9029-3af1904bcad2" alt=""><figcaption></figcaption></figure>

#### Assign permissions and finalize

1. Select `StoikAdditionalPermissions` and `SecurityAudit` as permissions.
2. Click `Next`.
3. Enter `stoik-auditor` as the role name.
4. Click `Create role`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-2794beb9987fd6c3fbc6d1381d782094e37a0f5c%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-29%20%C3%A0%2017.34.12.png?alt=media" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-8eb6ce6816a9efa007029ca8e39e3a2ba30a5bfd%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-29%20%C3%A0%2017.36.16.png?alt=media" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-89e73e7eac5001395f4b17220469b4dd16428864%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-29%20%C3%A0%2017.40.03.png?alt=media" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Add an ARN identifier

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F94ToUKwUHjDp3oDbMGwC%2FScreenshot%202026-04-07%20at%2018.51.37.png?alt=media&amp;token=a0b5b107-57e3-44a2-a094-164fac87050b" alt=""><figcaption></figcaption></figure>

1. Go to the `stoik-auditor` role you just created.
2. Copy the `ARN` reference.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-694a0c3196cb783df765b0e5e51364bfdedd2556%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-29%20%C3%A0%2017.44.21.png?alt=media" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-1f1c9950b3489ea4d1572e1acb6c4e87f34a3f33%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-29%20%C3%A0%2017.45.04.png?alt=media" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Complete setup

1. Paste the `ARN` reference into the dedicated field in the Stoïk Protect console.
2. Click `Complete set-up`.

{% hint style="info" %}
Refresh the Stoïk Protect page if necessary.
{% endhint %}
{% endstep %}
{% endstepper %}

## AWS Organizations

With AWS Organizations, create the `stoik-auditor` role in each member account you want scanned, not only in the management account. All the permissions requested are read-only: the AWS-managed `SecurityAudit` policy plus `StoikAdditionalPermissions`. There is no smaller set.

## Troubleshooting

<details>

<summary>"Integration with AWS failed" / "Unable to assume role"</summary>

Check these in order, most common first:

1. **External ID:** the value in Stoïk Protect must match the `sts:ExternalId` in the role's trust policy exactly. Case and spaces count. An external ID from an earlier attempt fails too.
2. **Trusted account:** the trust policy trusts account `436578779424`.
3. **Role:** `stoik-auditor` exists in that account, and the ARN pasted in Stoïk Protect is its ARN. For a role that does not exist, AWS answers "access denied", not "not found".
4. **Policies:** both `SecurityAudit` and `StoikAdditionalPermissions` are attached, and the latter matches the JSON above. Older setups often miss `sso:Describe*` and `sso:List*`.

</details>

<details>

<summary>Everything is correct, but the setup still says "roles not correctly configured"</summary>

This is usually a Service Control Policy (SCP). Before accepting an account, Stoïk checks its role with AWS's permission simulator, and some SCPs make that simulator refuse every permission, even though the real scan works. Contact <protect@stoik.io> with your AWS account ID.

</details>

<details>

<summary>The scan worked, then stopped</summary>

Something changed on your side: the role was deleted or renamed, a policy was detached, the external ID changed, or a new SCP was added. Go through the four checks in ["Integration with AWS failed" / "Unable to assume role"](#integration-with-aws-failed--unable-to-assume-role).

</details>


# Setup: GCP

Give the Cloud scan read-only access to a Google Cloud project by creating a service account that Stoïk can use.

## Prerequisites

* An administrator account is required to perform this deployment.
* A technical advisor is required for ongoing supervision and vulnerability remediation.
* One person does the whole setup, from start to finish: a Stoïk Protect user allowed to edit prevention tools, who is also an administrator of the GCP project.

## Setup

{% stepper %}
{% step %}

### Select Google Cloud Platform

Go to `Cloud Scan` > `Settings`, click `New Provider`, and select `Google Cloud Platform`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FDw2b46xxHQFAG4OCkigt%2FScreenshot%202026-04-07%20at%2018.57.56.png?alt=media&amp;token=72ccef3f-3cb8-4f54-b98b-6c748f5e3e92" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Open your Cloud Console

Go to the [Google Cloud Console](https://console.cloud.google.com/) and log in with your administrator account.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FUTcz9AktGTD9Cczz9KPB%2Fcloud-scan-google-cloud-console.png?alt=media&amp;token=4f6d18c5-3eec-4040-9d46-12ae365c6dd7" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Open the API library

Go to `APIs & Services` > `Library` in the left-hand menu.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FH7VRpz11JW2ppKogPQ4y%2Fcloud-scan-console-api-services-library.png?alt=media&amp;token=e0153ae0-5695-4746-b3bc-da256ee2d1bb" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Find the Cloud Resource Manager API

Search for `Cloud Resource Manager API` and click on it.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F4f8FTY2S563Lm31xK7Np%2Fcloud-scan-cloud-resource-manager-api.png?alt=media&amp;token=0aafc7a8-1d1d-40fc-b08d-b6eb064cb3d1" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Enable the API

Click `Enable`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FqHVrHGykLoST9s0PycrJ%2Fcloud-scan-resource-manager-api-enable.png?alt=media&amp;token=20c69a92-2c79-4568-a5c9-86c5806db366" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Open the service accounts

Go to `IAM & Admin` > `Service Accounts` in the left-hand menu.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FskzJLLEd9DokDw9pHIJF%2Fcloud-scan-iam-admin-service-accounts.png?alt=media&amp;token=c83333e7-9f3c-4fc9-9678-ff91daec5658" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Create the service account

Click `Create Service Account`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FBTbAgGb6fSK20BVXuuSz%2Fcloud-scan-create-service-account.png?alt=media&amp;token=168d4f6e-46ff-4420-a503-9002c33b3f2a" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Name the service account

Enter `stoik-scanner` as the service account name, then click `Create and continue`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F6whuJGKh4PNGgftwZqLV%2Fcloud-scan-create-service-account-name.png?alt=media&amp;token=d00f2170-bd90-455f-a9f3-94da30e44b62" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Assign roles

Add the following 3 roles: `Viewer`, `Security Reviewer`, `Stackdriver Accounts Viewer`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FCChwue8tVQSHKCpMIlCe%2Fcloud-scan-create-service-account-permissions.png?alt=media&amp;token=26b9ac91-e109-412d-8cd8-37548092f78b" alt=""><figcaption></figcaption></figure>

Leave step 3 empty and click `Done`.
{% endstep %}

{% step %}

### Open the service account

Click on the service account you just created, then click `Principals with access`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Ft4z2qlijQGNNDA4ISbLi%2Fcloud-scan-open-service-account.png?alt=media&amp;token=94f014b7-d982-4152-bfd9-8d3af5436de8" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FiCKjOoUMiZUKJDD9lcDz%2Fcloud-scan-service-account-grant-access.png?alt=media&amp;token=362b9a8b-d905-4d37-ad5f-b8017951ebe0" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Grant access

1. Select your service account and click `Grant access`.
2. Enter `stoik-scanner@stoik-gcp-scan-prod.iam.gserviceaccount.com` in the `New principals` field.
3. Assign the `Service Account Token Creator` role, then click `Save`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FtA2nkzIqHpx65ok06gqr%2Fcloud-scan-service-account-add-principals.png?alt=media&amp;token=f1c62da6-496e-4804-9afc-161d71df3c2b" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Copy the service account email

1. Click `Details` in GCP.
2. Copy the email address of the created service account.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FBbOUrFQXlScFI32CxLgR%2Fcloud-scan-service-account-copy-email.png?alt=media&amp;token=4fa17999-bb77-4764-bfff-99c2a20735f6" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Complete setup

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F6yEDTO51V6cbdB8McEz4%2FScreenshot%202026-04-07%20at%2018.59.05.png?alt=media&amp;token=1568ad28-5e33-4283-8b3c-d415d7606fce" alt=""><figcaption></figcaption></figure>

1. Paste the service account email address into the `Service Account Email` field in Stoïk Protect.
2. Click `Complete setup`.

{% hint style="info" %}
Refresh the Stoïk Protect page if necessary.
{% endhint %}
{% endstep %}
{% endstepper %}

## Projects to scan

Set up the projects that hold your servers and databases: a project created just for Stoïk is empty, so there is nothing to scan. The API must be enabled one project at a time.

## Troubleshooting

<details>

<summary>The setup failed, then worked when I retried</summary>

New roles, permissions and APIs take a few minutes to apply. Wait 5 minutes and click again.

</details>

<details>

<summary>"role/stackdriver.accounts.viewer must be assigned"</summary>

`Stackdriver Viewer` and `Stackdriver Accounts Viewer` are two different roles. The service account needs exactly `Viewer`, `Security Reviewer` and `Stackdriver Accounts Viewer`.

</details>

<details>

<summary>"do not belong to a permitted customer"</summary>

Your organization policy **Domain restricted sharing** (`iam.allowedPolicyMemberDomains`) blocks the Stoïk service account. Your admin adds Stoïk's customer ID `C01isw1kw` to the allowed values, in `IAM & Admin` > `Organization Policies`. This requires the `Organization Policy Administrator` role at the organization level: being Owner of the project is not enough.

</details>

<details>

<summary>"Integration with GCP failed" / getAccessToken denied</summary>

Two usual causes:

* The Cloud Resource Manager API is not enabled, or was enabled only a few minutes ago.
* `stoik-scanner@stoik-gcp-scan-prod.iam.gserviceaccount.com` did not get the `Service Account Token Creator` role on your service account.

</details>


# Reading results

The `Cloud Services` > `Results` dashboard shows the security level of your cloud infrastructure and the misconfigurations to address.

## Dashboard header

The top of the results page gives you an at-a-glance overview of your cloud security posture.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FrU4tdpQOozdQL1nfliEY%2FScreenshot%202026-04-07%20at%2018.18.59.png?alt=media&amp;token=21b0f43c-0df4-43b7-8bb6-ecb53ffaf8ac" alt=""><figcaption></figcaption></figure>

### Health status

The health status indicates the overall state of your cloud infrastructure security. It tells you immediately whether your configuration requires attention.

### Score

Your cloud security score is displayed out of 100. A score of 100 means no misconfigurations were detected. The lower the score, the more misconfigurations need to be addressed. The score is calculated based on the number and severity of detected misconfigurations across all your connected providers.

A score of 100 or empty results are usually correct: the scan ran and found nothing (small environment, or an Entra ID tenant with few users). The page may say "Results will be displayed here once the scan is complete" even though the scan is done: check the last scan date.

### Graph

The graph shows the evolution of your score over time, so you can track whether your cloud security posture is improving or degrading. Use it to monitor the impact of remediation actions you've taken.

## Misconfigurations

Only detected misconfigurations are displayed in the results table. Each misconfiguration shows:

* **Name:** the misconfiguration detected
* **Criticality:** Critical, High, Medium, or Low
* **Description:** an explanation of the issue and its potential impact

Click a misconfiguration to access the remediation steps required to resolve it.

Misconfigurations cannot be excluded, marked as accepted risk or marked as false positives today. This includes shared mailboxes, emergency accounts and users protected some other way: the Cloud scan shows everything it finds.

## MFA findings (Entra ID)

**Users with MFA Disabled** lists active accounts that have no MFA method registered in Entra ID. That is not the same as "MFA enforced": a user covered by a policy who has not registered a method yet, often an account that never signs in, still appears. The full list is in the finding's resources and in the export.

These accounts are not counted:

* accounts with sign-in blocked
* Entra Connect sync accounts
* guests, which have their own finding: `No MFA Policy for Guest Users`

If you enforce MFA with Conditional Access, the finding is dropped only if a policy is **On** (not report-only), includes **All users** (a group does not count), requires MFA (grant control or authentication strength), and can be read (`Policy.Read.All` granted). Otherwise, the listed users really have no MFA: register MFA for them, or cover them with an "All users" policy.

* **Emergency (break-glass) accounts:** they cannot be excluded from the finding. Excluding them from your "All users" MFA policy is fine: the policy still counts and the finding disappears.
* **Shared mailboxes and meeting rooms:** they appear because they can still sign in. Block sign-in for them in Microsoft 365 (Microsoft recommends this for shared mailboxes): they disappear at the next scan.

To check MFA status per user in Azure, see [MFA status](https://docs.stoik.io/prevention-tools/cloud-scan/azure/mfa-status).

## Other findings

* **Shadow Admin Users:** accounts that hold one of these Entra roles without being Global Administrator: Application, Cloud Application, Authentication, Privileged Authentication, Privileged Role, Password, Helpdesk or User Administrator. These roles let someone become admin, so dedicated admin accounts are flagged on purpose.
* **High Percentage of Privileged Users:** only appears when there are at least 4 privileged accounts and they are more than 10% of all users.
* **Root Account Used Recently (AWS):** the AWS root account was used to sign in during the last 90 days.

## Exporting results

Export the results table in CSV or XLS format using the `...` button at the right of the search bar.

{% hint style="info" %}
To fix a finding, see the [Cloud Scan remediation guide](https://docs.stoik.io/cyber-best-practices/vulnerability-remediation-guides/cloud-scan) for step-by-step fixes per provider (Microsoft Azure / Entra ID / Microsoft 365, AWS, GCP).
{% endhint %}

## Troubleshooting

<details>

<summary>A fixed misconfiguration is still reported</summary>

* **Has the scan run since?** See [The results haven't updated](#the-results-havent-updated).
* **Can Stoïk see the fix?** Conditional Access changes are invisible without `Policy.Read.All`, see [Policy.Read.All cannot be granted](https://docs.stoik.io/pages/o078MxFZGrJa3CHRphbh#policy.read.all-cannot-be-granted-or-a-permission-is-missing).
* **Does the fix match what is checked?** For example, a Conditional Access policy for a group of users does not count as "All users", and a custom role that nobody is assigned to does not count.
* **Is it a known limitation?** These findings can stay even when the setting is in place, and you can ignore them:
  * GCP `Log Metric Filter Doesn't Exist…` / `Alerts Doesn't Exist For…`
  * Azure `No Administering Resource Locks Role` when the role is given through PIM (eligible assignment): only permanent assignments are seen.

</details>

<details>

<summary>The results haven't updated</summary>

Check the last scan date, then:

* **Stoïk's access was removed or changed.** Before each scan, Stoïk checks that it still has access. After 3 failed checks in a row, the scans of that provider stop. Once the access is fixed, do the setup again to restart the scans.
* **The environment is very large.** A scan stops after 1 hour, and very large environments can hit that limit.

Restart the scan from `Cloud Scan` > `Settings`. If the date still has not changed the next day, contact <protect@stoik.io>.

</details>


# Credentials

The `Credentials` section gives you a dedicated view of everything Stoïk has found related to leaked credentials and compromised devices for your organisation.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-database">:database:</i></h4></td><td><h4>Exposed Data</h4></td><td>Browse every leak found about your organisation.</td><td><a href="https://docs.stoik.io/prevention-tools/credentials/exposed-data">Exposed Data</a></td></tr><tr><td><h4><i class="fa-key">:key:</i></h4></td><td><h4>Valid Credentials</h4></td><td>Fix a leaked password that still works.</td><td><a href="https://docs.stoik.io/prevention-tools/credentials/valid-credentials">Valid Credentials</a></td></tr><tr><td><h4><i class="fa-user-lock">:user-lock:</i></h4></td><td><h4>Leaked password</h4></td><td>Why Stoïk does not share it, and what to do instead.</td><td><a href="https://docs.stoik.io/prevention-tools/credentials/leaked-password">Leaked password</a></td></tr><tr><td><h4><i class="fa-laptop">:laptop:</i></h4></td><td><h4>Infected devices</h4></td><td>Handle a device compromised by infostealer malware.</td><td><a href="https://docs.stoik.io/prevention-tools/credentials/infected-devices">Infected devices</a></td></tr></tbody></table>

## How it works

Stoïk Protect is connected to three external intelligence providers:

* **LeakCheck** monitors public data breaches: incidents where a company's user database was hacked and dumped online.
* **Hudson Rock** and **Flare** collect infostealer logs: data silently exfiltrated from employee devices infected by malware.

From the assets you've declared in the [External Scan](https://docs.stoik.io/prevention-tools/external-scan/configuration) (domain names, IP addresses), Stoïk continuously queries these providers for anything related to your organisation. Results flow into two tabs: **Exposed Data** and **Alerts**.

## What Stoïk looks for

* Every email address on a domain declared in the External Scan, or found by it, including domains that redirect to yours.
* Not personal addresses (Gmail and others), even when an employee uses them for work.
* For prospects, Stoïk collects leaks but does not test passwords: Valid Credentials alerts only exist for customers.

## Exposed Data

Exposed Data is the full intelligence feed: every piece of data Stoïk has found leaked about your organisation online. An entry does not necessarily mean the credential still works today. See [Exposed Data](https://docs.stoik.io/prevention-tools/credentials/exposed-data).

## Alerts

Alerts are what Stoïk has actively confirmed needs immediate action. There are two types:

* **Valid Credentials:** a leaked login and password that still grants access to one of your services. See [Valid Credentials](https://docs.stoik.io/prevention-tools/credentials/valid-credentials).
* **Infected Device:** an employee device identified as compromised by infostealer malware. See [Infected devices](https://docs.stoik.io/prevention-tools/credentials/infected-devices).

## How Alerts and Exposed Data are linked

Valid Credentials alerts are built from Exposed Data: Stoïk picks leaked credentials from the feed and tests them against your services. Those that still work are promoted to Alerts.

Infected Device alerts work in the other direction: each compromised device typically produces a payload of stolen data (passwords, cookies, browser data) that flows into Exposed Data. Opening an Infected Device alert shows you the associated Exposed Data entries.

## Data refresh frequency

Stoïk queries each provider on a regular schedule:

* Flare: every 24 hours
* Hudson Rock: every 72 hours
* LeakCheck: every 7 days

Exposed Data and Infected Device alerts follow the same schedule as their respective provider.

**Valid Credentials alerts** are tested as soon as new exposed credentials are detected. Stoïk then retests them every 3 days until the credential no longer works, so the alert automatically resolves once the password has been changed. Marking the alert as fixed in Stoïk Protect triggers a new test sooner.

## Alert emails

Alert emails are not automatic: the CERT reviews each alert and sends the email by hand, which can take a few days. An alert marked `Waiting on customer` means the email was sent: check your spam folder.


# Exposed Data

Exposed Data is the full intelligence feed: every piece of data Stoïk has found leaked about your organisation online.

## Sub-tabs

Exposed Data is organised into three sub-tabs:

* **Raw logs:** every leak record, all data types (credentials, cookies, emails, phone numbers, IP addresses, …).
* **Credentials:** credential-type leaks only: URL + login + password.
* **Cookies:** session cookies exfiltrated from infected devices.

For each entry, the side panel shows the intelligence provider that found it, the estimated leak date, the data type, and any linked alert.

Review new entries and acknowledge them once investigated. Acknowledged entries remain visible but no longer count as new.

{% hint style="info" %}
An entry in Exposed Data does not necessarily mean the credential still works today. If we have confirmed it does, it appears as a [Valid Credentials](https://docs.stoik.io/prevention-tools/credentials/valid-credentials) alert in the Alerts tab.
{% endhint %}

## Reading an entry

* **Domain:** the website the credentials give access to. `microsoftonline.com` means a Microsoft account. On older Flare entries it is the domain Stoïk searched instead, because Flare does not say which site the leak came from.
* **Password and Encrypted password:** when both are listed, only an encrypted (hashed) password leaked. There is no clear-text password.

## Troubleshooting

<details>

<summary>I don't recognise an email address</summary>

The usual explanations:

* **A former employee:** the leak is older than the deletion of the mailbox.
* **Someone signed up on a website with your domain without owning the address:** many sites never check the email. If that account works, make a discard request and ask the site to stop using your domain.
* **Fake data:** sellers pad leaks to make them look bigger. Typical signs are subdomains that don't exist (`@www.client.fr`, `@pro.client.fr`) and passwords that all follow the same pattern. Mark the entries as Acknowledged.

</details>

<details>

<summary>I don't recognise a domain</summary>

It was usually added to the External Scan and later removed, or it redirects to one of your domains. After each scan, domains the External Scan no longer finds are removed from Credentials, with their results: restart the scan.

If the domain is still there once the scan has finished, or you only excluded it, contact <protect@stoik.io>.

</details>

<details>

<summary>A leak is missing in Stoïk Protect</summary>

* The Valid Credentials tab only lists passwords Stoïk confirmed still work. Every leak is under `Exposed Data` > `Raw logs`.
* Exposed Data only shows new entries by default. If a page says no cookies were found, or "show N more logins" opens an empty list, tick **Acknowledged** in the filters.

If it is still missing, contact <protect@stoik.io> with the email address concerned.

</details>


# Valid Credentials

A Valid Credentials alert means Stoïk found a leaked URL + login + password in [Exposed Data](https://docs.stoik.io/prevention-tools/credentials/exposed-data) and **confirmed it still grants access** to one of your services today (Microsoft 365, Fortinet VPN, Citrix, exposed web panels, and others).

This is not a historical leak. It is a working credential that an attacker could use right now.

## How Stoïk tests a password

Stoïk makes one automated login attempt, chosen from where the password leaked: Microsoft 365, a Fortinet VPN, a Citrix gateway, or the website in the leak (a browser agent fills in the login form). Each alert shows the method. Stoïk never browses inside the account.

The tests come from an AWS IP range located in Ireland, with no fixed IP, like the [External Scan](https://docs.stoik.io/prevention-tools/external-scan). Sign-ins from Ireland on leaked accounts in your logs are probably Stoïk's. If you need the exact time and IP of the attempts, contact <protect@stoik.io>.

## Resolving an alert

1. Reset the password for the affected account immediately.
2. Enable multi-factor authentication (MFA) on that account if not already active.
3. Check whether the same password is reused on other services. Change those too.
4. Mark the alert as fixed once done. This triggers a new test sooner: otherwise, Stoïk retests every 3 days.

Stoïk does not share the leaked password. See [Leaked password](https://docs.stoik.io/prevention-tools/credentials/leaked-password) for the reasons and for what to do when the employee is gone.

## Troubleshooting

<details>

<summary>MFA is enabled, so why is there an alert?</summary>

Stoïk raises the alert as soon as the service accepts the password, even if MFA comes next. The password is still compromised and only MFA protects the account: the fix is still a reset.

</details>

<details>

<summary>I reset the password, but the alert is still there</summary>

Stoïk retests every 3 days while the password still works. Marking the alert as fixed triggers a new test sooner. Then check the last tested date on the alert:

* **After the reset, and still valid:** the reset didn't reach the account Stoïk tests. It is often another account with the same email (another tenant, a third-party or personal account). Don't mark it as a false positive.
* **Weeks old:** the retest may be stuck. Contact <protect@stoik.io>.

</details>

<details>

<summary>The account doesn't exist</summary>

The alert shows where Stoïk tested the password and how. It is often an account the employee created with their work email outside your company's systems, for example a personal Tableau ID instead of the company's Tableau Cloud site. Your admin console doesn't show it, but the password works there.

The answer is the same: a reset through the email address, as described in [Leaked password](https://docs.stoik.io/prevention-tools/credentials/leaked-password). If the tested URL looks like a signup page, or you think the test is wrong, contact <protect@stoik.io> with the alert link.

</details>

<details>

<summary>Why was an alert archived?</summary>

A retest found that the password no longer works, or the service (a VPN, for example) has been unreachable for a week. Stoïk Protect doesn't show the reason or the test history yet.

</details>

<details>

<summary>The Exclude button is greyed out</summary>

While the alert is in `CERT Analysis`, you can only make a discard request from the side panel. A password that still works is not a false positive, even if the employee has left.

</details>


# Leaked password

Stoïk Protect shows only the first and last characters of a leaked password, on purpose. This page explains why, and what to do instead.

## Why Stoïk does not share the password

* **The fix is a reset, not sharing.** Knowing the password doesn't reduce the exposure. Changing it does.
* **The password belongs to the employee, not to the company.** People reuse their passwords: this one may also open the person's personal mailbox or social networks. Handing it to their employer, or former employer, may give them the key to the person's private life, while Stoïk knows nothing about the context: a dispute, a dismissal, a conflictual resignation.
* **You don't need more to act.** The alert shows the login and the service where the password works.

## What Stoïk does and doesn't do

Stoïk does not log in manually to an account, change its password or close it. Our automated test is a single login attempt, without browsing inside the account. A manual login would add nothing to it, and we never access an account without its owner's consent.

Changing the password or closing the account ourselves would mean changing an account on someone else's service and locking its owner out, when you can get the same result yourself.

## What to do instead

In this order:

1. Ask the employee, or former employee, to change the password or close the account.
2. Otherwise, go through the domain: restore the mailbox or create a temporary alias for the address, then use "Forgot password" on the service.
3. Otherwise, contact the service's support with proof that you own the domain.

If you don't want to do anything, set the alert to Acknowledged and accept the risk.

{% hint style="warning" %}
Disabling the employee's mailbox isn't enough: the account on the other service still works for anyone who has the password.
{% endhint %}


# Infected devices

An Infected Device alert means one of your employee devices has been identified as compromised by infostealer malware, by Hudson Rock or Flare. The device is actively exfiltrating data (passwords, cookies, browser autofill, saved credentials) until it is fully wiped.

## Resolving an alert

1. Isolate the device from the network immediately (disconnect from Wi-Fi and LAN).
2. Wipe and reinstall the device from scratch. A virus scan alone is not sufficient.
3. Force a password reset for all accounts used on that device.
4. Invalidate active sessions on those accounts (sign out everywhere) to neutralise any stolen session cookies.
5. Enable MFA on the affected accounts.

## Personal devices

Stoïk detects infected devices based on professional account activity. It cannot determine whether the device itself is a company-issued machine or a personal one. A personal device (a phone or laptop on which an employee connected a professional account) can therefore appear as an Infected Device alert. If you identify an alert as relating to a personal device, update the device type from the alert side panel:

1. Open the alert.
2. Switch the **Device type** from *Professional* to *Personal*.

The alert moves to **Dismissed** status. It remains visible in the list but is filtered out by default and no longer counts as an active alert.

{% hint style="info" %}
Even on a personal device, the professional accounts used on it remain at risk. Before dismissing the alert, make sure the employee has reset their passwords and invalidated active sessions for any professional account they accessed on that device.
{% endhint %}

{% hint style="warning" %}
If you believe an alert is a false positive, you can request a discard directly from the side panel. Stoïk will review the request before closing it.
{% endhint %}

## Troubleshooting

<details>

<summary>There is no device name, or I don't recognise the device</summary>

The alert is real: the malware didn't capture those details, or the host name it reported means nothing (for example "PRIVAT"). A device you don't recognise is usually a personal one where an employee used a work account: use the username or email in the leaked data to find them.

Stoïk doesn't tell personal and work devices apart: you change the device type yourself, see [Personal devices](#personal-devices). If the alert shows no user at all, contact <protect@stoik.io>: the log often holds a username that Stoïk Protect doesn't display.

</details>

<details>

<summary>The device was cleaned long ago</summary>

Providers buy stealer logs as snapshots, so they can't know when a device was wiped. An infected device is archived automatically 180 days after its last leak. For many old devices at once, contact <protect@stoik.io>: the CERT can close them in bulk as patched.

</details>

<details>

<summary>I had an infection years ago and don't see it</summary>

Providers keep everything they collect, but they focus on recent leaks: an old infection may never show up.

</details>


# Deploy the EDR

Check the deployment rules and supported hardware, then deploy Stoïk MDR (EDR) from Stoïk Protect or by following the guide for your EDR and operating system.

## Prerequisites

* Uninstall any existing antivirus software before deployment. To avoid conflicts between EDR and any existing antivirus software, we recommend uninstalling any antivirus or similar solutions before installation. See [Antivirus and exclusions](https://docs.stoik.io/mdr/deploy/antivirus).
* Your equipment is supported: see [Supported environments](#supported-environments).

## Basic rules for deployment

* **No restart required**. Neither your servers nor user workstations need to be restarted. EDR deployment is simple and requires no service interruption.
* **Immediate internet connection not required.** Endpoints can be installed offline, but will need to connect to the internet later to fully activate the EDR.
  * If deploying via MDM or GPO, add `ProvNoWait=1` to the install command: `WindowsSensor.exe /install /quiet /norestart CID=<customer_ID> ProvNoWait=1`

## Two ways to deploy

1. **From the Stoïk Protect app (recommended).** Available environments depend on the provider:

   * CrowdStrike: Windows, macOS, Linux, ChromeOS, Virtual environments, and Mobile
   * SentinelOne: Windows, macOS, Linux

   The screenshot below shows the view if your provider is CrowdStrike:

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F78rjVLESH4LyyVG5pli0%2Fimage.png?alt=media&amp;token=f60eee67-d7f0-481b-8ced-d5cc271bad74" alt=""><figcaption></figcaption></figure>
2. **From the Help Center documentation.** Follow the guide for your EDR and operating system:
   * CrowdStrike: [Windows](https://docs.stoik.io/mdr/deploy/crowdstrike/os/windows), [macOS](https://docs.stoik.io/mdr/deploy/crowdstrike/os/macos), [Linux](https://docs.stoik.io/mdr/deploy/crowdstrike/os/linux), [ChromeOS](https://docs.stoik.io/mdr/deploy/crowdstrike/os/chromeos), [Virtual environments](https://docs.stoik.io/mdr/deploy/crowdstrike/os/virtual-environments), or [Mobile devices](https://docs.stoik.io/mdr/deploy/crowdstrike/add-on/mobile)
   * SentinelOne: [Windows](https://github.com/stoikio/gitbook-global/tree/main/help-center/mdr/deploy/sentinelone/os/windows.md), [macOS](https://docs.stoik.io/mdr/deploy/sentinelone/os/macos), or [Linux](https://docs.stoik.io/mdr/deploy/sentinelone/os/linux)
   * Microsoft Defender: see [Microsoft Defender](https://docs.stoik.io/mdr/deploy/microsoft-defender).

## Supported environments

Stoïk MDR is designed for **IT workstations and servers only**. It cannot be installed on industrial or embedded hardware. If you are unsure whether your equipment is compatible, contact us at <protect@stoik.io>.

The following hardware **is supported:**

* Standard x86\_64 (Intel / AMD) or server-grade ARM64 workstations and servers
* Windows, macOS, Linux with standard vendor kernels (Debian, Ubuntu, RHEL, etc.)

For detailed system requirements, refer to your vendor's documentation.

The following hardware **is not supported:**

* **OT / IoT devices**: industrial gateways, PLCs, SCADA controllers
* **Embedded systems**: hardware from Compulab, Advantech, Moxa, Wago, Siemens Industrial, etc.
* **Custom or stripped-down kernels**: missing standard modules such as `CONFIG_DMI_SYSFS`
* **Low-resource embedded SoCs**: Cortex-A53, i.MX8, and similar processors with limited RAM / storage


# Antivirus and exclusions

Uninstall any existing antivirus or other EDR solution before installing your Stoïk MDR EDR (CrowdStrike or SentinelOne), to prevent conflicts that could compromise the proper functioning of the solution.

## When to uninstall

{% tabs %}
{% tab title="CrowdStrike" %}
The Stoïk teams have implemented a gradual [three-phase deployment](https://docs.stoik.io/mdr/deploy/crowdstrike#deployment-phases) to limit the risk of blockages and the impact on user workstations. Phase 1 lets you install the EDR while keeping your antivirus or EDR in place, so you must uninstall your previous antivirus or EDR solution before moving on to phase 2.

{% hint style="warning" %}
If you do not uninstall it before moving on to phase 2, a "blue screen" may appear when the workstation is restarted.
{% endhint %}

However, some antivirus or EDR solutions, whether single-use or multi-use, may cause blockages as early as phase 1 of the deployment: see [Single-use antivirus](#single-use-antivirus) and [Multi-purpose solutions](#multi-purpose-solutions).
{% endtab %}

{% tab title="SentinelOne" %}
SentinelOne does not use a phased deployment: full protection is active as soon as it is installed. Uninstall your previous antivirus or EDR solution before installing SentinelOne, to avoid any conflict.
{% endtab %}
{% endtabs %}

## Single-use antivirus

These programs have only one function: antivirus protection.

{% tabs %}
{% tab title="CrowdStrike" %}

* **Bitdefender**: must be uninstalled during phase 1 of deployment. If you do not uninstall it, a blue screen may occur when you restart your computer.
* **Trend Micro**: must be uninstalled starting from deployment phase 2, otherwise error messages may appear.
* **macOS**: No known conflicts, no action required.
  {% endtab %}

{% tab title="SentinelOne" %}

* **Bitdefender** and **Trend Micro**: must be uninstalled before installation, since SentinelOne has no phased deployment.
* **macOS**: No known conflicts, no action required.
  {% endtab %}
  {% endtabs %}

## Multi-purpose solutions

These software programs offer several security functions in a single tool: antivirus, firewall, USB device control, integrated EDR, etc. Common examples include **WithSecure**, **ESET**, and **Kaspersky**.

{% tabs %}
{% tab title="CrowdStrike" %}
Disable the antivirus functionality only. There is no need to uninstall the entire software.

There is no need to uninstall **Windows Defender**: it is fully compatible with your EDR. It remains active in phase 1, and the antivirus engine is disabled automatically at phase 2.
{% endtab %}

{% tab title="SentinelOne" %}
These solutions must be fully uninstalled before installing SentinelOne, since there is no phased deployment to tolerate a temporary conflict.

The only exception is **Windows Defender**: there is no need to uninstall it, it is fully compatible with SentinelOne, and the antivirus engine is disabled automatically as soon as SentinelOne is installed.
{% endtab %}
{% endtabs %}

{% hint style="warning" %}
Beware of side effects on Windows. If you completely disable the multi-purpose solution, and it previously managed your firewall, then:

* The [Windows Firewall](https://learn.microsoft.com/en-us/windows/security/operating-system-security/network-security/windows-firewall/) will automatically take over.
* However, it will apply its own default, very restrictive rules.
* This can completely block your network connections.

The solution is to reconfigure your local Windows firewall with the necessary rules and exclusions to restore the expected connections.
{% endhint %}

Example of an error message:

<figure><img src="https://25674881.fs1.hubspotusercontent-eu1.net/hubfs/25674881/signed-url-redirect%20(1).png" alt=""><figcaption></figcaption></figure>

## Exclusion rules

Managing application flows through exclusion rules is an approach specific to traditional antivirus solutions, but it does **not** apply to an EDR such as **CrowdStrike** or **SentinelOne**.

### Difference in operation between antivirus and EDR

* **Antivirus:** proactively scans all dormant files on the system (whether executed or not). This generates a large volume of scans and requires numerous exclusions to avoid false positives on inactive files.
* **EDR (CrowdStrike, SentinelOne):** focuses only on files that are actually executed and on dynamic system behaviors (processes, connections, user actions, etc.). No continuous full-disk scanning is performed. Inactive files are not analyzed except during manual scans.

Since an EDR analyzes **suspicious behaviors rather than signatures**, excluding a folder from analysis is unnecessary.

{% hint style="warning" %}
Excluding a folder would have a negative effect: if suspicious behavior were to occur in that folder, the EDR would not be able to detect it. That said, this remains technically possible if the need is validated by our SOC team.
{% endhint %}

### Intelligent prevention and reactive exclusion management

Your EDR integrates a set of preventive mechanisms to limit false positives without manual intervention:

* Preconfigured exclusion lists are automatically applied during deployment.
* In the event of a suspicious detection involving a business application, the program is temporarily blocked, then analyzed by our SOC before any definitive action is taken.

{% hint style="warning" %}
To avoid potential unwanted business disruptions, full isolation is performed **only manually**, either after analysis by our SOC team or by you.
{% endhint %}

### Exclusions should not be defined upfront

Trying to anticipate exclusions before any detection occurs represents a major risk:

* It amounts to preemptively excluding potentially malicious behaviors simply because they are associated with a known business application.
* **Result:** you risk neutralizing the EDR's detection capabilities, allowing certain attacks to go undetected.

Exclusions in an EDR are far fewer. At **Stoïk**, they are handled **only in response to a real detection**, when a blocking event has actually occurred, based on SOC analysis. It is neither necessary nor advisable to add "preventive" exclusion rules as you would with a traditional antivirus.

**Concrete example:** One of our clients (a national retail player) had more than 3,000 workstations and several thousand exclusions in their previous antivirus solution. After switching to an EDR, no exclusions were required.


# CrowdStrike

Once the EDR is set up on a workstation or server, monitoring starts automatically and the level of protection increases gradually. No action is required on your part unless otherwise requested.

## Deployment phases

The EDR ramps up in three successive phases over approximately one month, under the supervision of Stoïk's cybersecurity engineers.

### Phase 1: Detection (observation mode)

* The EDR works like an intelligent antivirus without disrupting activity.
* It monitors system activity (files, processes, connections, etc.) and collects all relevant logs.
* It only blocks critical threats (ransomware, major malware).

{% hint style="info" %}
For workstations only, Phase 1 is skipped and endpoints are set up directly at Phase 2.
{% endhint %}

### Phase 2: Enhanced prevention

* The first preventive actions are put in place and certain suspicious activities begin to be blocked automatically.
* Exclusion rules are added if needed to ensure compatibility with your business software.
* The quarantine engine is activated, which may cause conflicts with existing antivirus software: see [Antivirus and exclusions](https://docs.stoik.io/mdr/deploy/antivirus).
* For customers with Windows Defender, this module is automatically disabled.

{% hint style="info" %}
CrowdStrike's Falcon Sensor may not appear in some fleet management tools such as Windows Security before Phase 2 is activated.
{% endhint %}

### Phase 3: Optimal protection

* The EDR blocks all malicious activity in real time. The security level is at its maximum and the IT system is proactively protected.
* The gradual ramp-up phases apply to the EDR module only. Other modules are automatically at maximum detection capacity from the start.

{% hint style="info" %}
Need to speed up or slow down the ramp-up? Contact us at <protect@stoik.io> and we can adapt the pace to your requirements.
{% endhint %}

## Monitored assets appear automatically in Stoïk Protect

All protected endpoints appear automatically in `MDR` > `Monitored Assets` in Stoïk Protect, on average 5 to 10 minutes after installation.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-0f3fd4b18a821c491cd0d38912c289ea2a0b9bf9%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2015.33.07.png?alt=media" alt=""><figcaption></figcaption></figure>

If a workstation is renamed, its name is automatically updated in the monitored assets list. It is not possible to rename a workstation directly from the list.


# Operating systems

Install the CrowdStrike EDR on each operating system in your environment.

{% hint style="info" %}
A simplified, guided installation walkthrough is also available directly in Stoïk Protect, under `MDR` > `Settings`. You can still follow the step-by-step guides below, which include screenshots.
{% endhint %}

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-windows">:windows:</i></h4></td><td><h4>Windows</h4></td><td>Install the EDR manually, from the console, by GPO or through an MDM.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/os/windows">Windows</a></td></tr><tr><td><h4><i class="fa-apple">:apple:</i></h4></td><td><h4>macOS</h4></td><td>Install and activate the Falcon sensor manually or through an MDM.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/os/macos">macOS</a></td></tr><tr><td><h4><i class="fa-chrome">:chrome:</i></h4></td><td><h4>ChromeOS</h4></td><td>Connect your ChromeOS devices through the Google Admin console.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/os/chromeos">ChromeOS</a></td></tr><tr><td><h4><i class="fa-linux">:linux:</i></h4></td><td><h4>Linux</h4></td><td>Install the RPM, DEB or SUSE package and activate the sensor.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/os/linux">Linux</a></td></tr><tr><td><h4><i class="fa-server">:server:</i></h4></td><td><h4>Virtual environments</h4></td><td>Install the sensor on a golden image for persistent or non-persistent VMs.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/os/virtual-environments">Virtual environments</a></td></tr><tr><td><h4><i class="fa-triangle-exclamation">:triangle-exclamation:</i></h4></td><td><h4>Installation errors</h4></td><td>Fix the most common errors during installation.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/os/errors">Installation errors</a></td></tr></tbody></table>


# Windows

Install the CrowdStrike EDR on Windows endpoints manually, from the console, by GPO or through an MDM. For a large-scale deployment, we recommend starting with a manual test installation on a few endpoints.

## Prerequisites

* Your current antivirus is completely uninstalled: CrowdStrike replaces it. If the firewall is managed separately, keep it running. If it is part of the same solution, copy any custom rules to Windows Firewall first. See [Antivirus and exclusions](https://docs.stoik.io/mdr/deploy/antivirus).
* Windows 7 or above. For older systems, see [Legacy systems](#legacy-systems).

## Install the EDR

{% stepper %}
{% step %}

### Download the installer

The executable is only available from Stoïk Protect: go to `MDR` > `Settings`, click `Windows`, then click `Download the executable`.
{% endstep %}

{% step %}

### Find your customer ID

Go to `MDR` > `Settings` and copy your customer ID.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FUFFCuuQGmvnkhwDCjtCX%2Fimage.png?alt=media&amp;token=880090b5-c3f0-4b45-9f75-fd4b6cc0bc4c" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Install CrowdStrike

{% tabs %}
{% tab title="Manual" %}

1. Run the executable on the endpoint.
2. Enter your customer ID in the `Customer ID with Checksum` field.
3. Click `Install`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FkFnum8LatH6UWOYhaYvq%2Funnamed.png?alt=media&amp;token=d9b09b2b-fe55-40b6-92e8-c24b638d552b" alt=""><figcaption></figcaption></figure>
{% endtab %}

{% tab title="Console" %}
Run the following command, replacing `<customer_ID>` with your customer ID:

```
<file_name>.exe /install /quiet /norestart CID=<customer_ID>
```

<details>

<summary>Special cases</summary>

**Adding tags**

To organize installations by site, country, or entity, add the `GROUPING_TAGS` option. Separate multiple tags with commas:

```
<file_name>.exe /install /quiet /norestart CID=<customer_ID> GROUPING_TAGS="tag1,tag2"
```

**Temporarily offline machine**

Add `ProvNoWait=1` to install without waiting for cloud connectivity. CrowdStrike still requires an internet connection to function and send alerts.

```
<file_name>.exe /install /quiet /norestart CID=<customer_ID> ProvNoWait=1
```

</details>
{% endtab %}

{% tab title="GPO" %}

1. Place the executable in a network share accessible to all workstations. Note the path, e.g. `\\ServerName\SharedFolder\CrowdStrike<file_name>.exe`.
2. Create an installation script named `InstallCrowdStrike.bat` with the following command, replacing `<customer_ID>` with your customer ID:

   ```
   \\ServerName\SharedFolder\CrowdStrike<file_name>.exe /install /quiet /norestart CID=<customer_ID>
   ```
3. Create a GPO:
   1. Open the Group Policy Management Console: `Windows + R` > `gpmc.msc` > `Enter`.
   2. Right-click the target container and click `Create a Group Policy Object in this domain and link it here`. Name it (e.g. `CrowdStrike Deployment`).
   3. Copy the script to `\\<server_name>\SysVol\<domain_name>\Policies\<GPO_GUID>\Machine\Scripts\Startup`.
   4. Right-click the new GPO and select `Edit`.
   5. Go to `Computer Configuration` > `Policies` > `Windows Settings` > `Scripts (Startup/Shutdown)`.
   6. Double-click `Startup`, click `Add`, then `Browse` and select `InstallCrowdStrike.bat`.
4. Apply the GPO:
   1. Run `gpupdate /force` on a domain controller.
   2. Restart a target machine to trigger the script.
   3. Run `gpresult /h gpresults.html` and verify the GPO is present on the machine.
   4. Verify that CrowdStrike is installed and operational.
   5. If you encounter any issues, check `C:\Windows\Debug\StartupLog.txt`.
      {% endtab %}

{% tab title="MDM" %}
**Intune**

1. Download the required components:
   * [Win32 Content Prep Tool](https://github.com/microsoft/Microsoft-Win32-Content-Prep-Tool)
   * CrowdStrike executable from Stoïk Protect: `MDR` > `Settings` > `Windows` > `Download the executable`
2. Open the Win32 Content Prep Tool, add the `.exe` file and enter the install command:

   ```
   <file_name>.exe /install /quiet /norestart CID=<customer_ID>
   ```
3. Download the generated `.intunewin` file.
4. Log in to [Intune Admin Center](https://intune.microsoft.com/), import the `.intunewin` file and assign it to your computer groups.

**NinjaOne**

Follow the [CrowdStrike with NinjaOne integration guide](https://www.ninjaone.com/docs/integrations/how-to-connect-crowdstrike/).
{% endtab %}
{% endtabs %}
{% endstep %}
{% endstepper %}

## Verify the installation

To verify the EDR installation on a machine, run:

```
sc query csagent
sc query csfalconservice
```

## Network traffic

If outbound traffic is blocked, see [Network requirements](https://docs.stoik.io/mdr/deploy/crowdstrike/network) for the addresses to allow.

## Legacy systems

The standard installer requires Windows 7 or above. For older systems, request the installer at <protect@stoik.io>:

* Windows XP 32-bit SP3 / 64-bit SP2
* Windows Server 2003 32/64-bit SP2
* Windows Server 2003 R2 32/64-bit SP2
* Windows Vista 32/64-bit SP2
* Windows Server 2008 32/64-bit SP2
* Windows Embedded POSReady 2009
* Windows 8 / 8.1 32/64-bit


# macOS

Install the CrowdStrike EDR on macOS endpoints manually or through an MDM. For a large-scale deployment, we recommend starting with a manual test installation on a few endpoints.

{% stepper %}
{% step %}

## Download the installer

The executable is only available from Stoïk Protect: go to `MDR` > `Settings`, click `macOS`, then click `Download the executable`.
{% endstep %}

{% step %}

## Find your customer ID

Go to `MDR` > `Settings` and copy your customer ID.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FUFFCuuQGmvnkhwDCjtCX%2Fimage.png?alt=media&amp;token=880090b5-c3f0-4b45-9f75-fd4b6cc0bc4c" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

## Install CrowdStrike

{% tabs %}
{% tab title="Manual" %}

#### Install the package

1. Launch `FalconSensorMacOS.LionLanner.pkg`.
2. Click `Continue` twice.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-f907a4be295266d78e20c4ac138689153236288c%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2016.02.01.png?alt=media" alt=""><figcaption></figcaption></figure>

3. Click `Install`.

#### Activate the sensor

A graphical interface opens automatically.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-32271e0eef04578b6ebf1aa585ca5f086ad07dcd%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2016.05.57.png?alt=media" alt=""><figcaption></figcaption></figure>

1. Enter your customer ID and click `Continue`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-4b01f734d3e65979212712aaae560deadd9f8422%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2016.24.28%20(1).png?alt=media" alt=""><figcaption></figcaption></figure>

2. Click `Setup`, then `Allow` on the permission dialog, then `Continue`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-2885979fd04a0e4217455e9caeb9fc26840a5911%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2016.36.40.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
If the sysextd error appears, restart your computer.
{% endhint %}

3. Click `Setup`, then `Open System Settings`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fk3dyQpvxN1TxNRfrbsBM%2Fimage.png?alt=media&amp;token=4f8afdfc-e966-4998-950e-667f5ce698e3" alt=""><figcaption></figcaption></figure>

4. Check the box and click `Done`.
5. Go back to the setup window and click `Continue`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-5566a0acd9e4d1c47642f3be5140b3fb84543b60%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2017.22.11.png?alt=media" alt=""><figcaption></figcaption></figure>

6. Go to `System Settings` > `Privacy & Security` > `Full Disk Access`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-902b1134ccb55097d61a816d2b949c8069ebc17d%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2016.37.54.png?alt=media" alt=""><figcaption></figcaption></figure>

7. Make sure the `Falcon Sensor` checkbox is checked.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-5e8ad4319e88e58e39495c0c1e32f6f800418abc%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-12%20%C3%A0%2016.38.08.png?alt=media" alt=""><figcaption></figcaption></figure>
{% endtab %}

{% tab title="MDM" %}

1. Install `FalconSensorMacOS.LionLanner.pkg` on the endpoint.
2. Run the following command, replacing `<customer_ID>` with your customer ID:

```bash
sudo /Applications/Falcon.app/Contents/Resources/falconctl license <customer_ID>
```

The EDR is now activated with your CID.
{% endtab %}
{% endtabs %}
{% endstep %}
{% endstepper %}


# ChromeOS

Connect your ChromeOS devices to the CrowdStrike EDR from the Google Admin console.

## Prerequisites

* Your ChromeOS hostname and API key. Request them from Stoïk at <protect@stoik.io>.

{% stepper %}
{% step %}

## Add the CrowdStrike XDR provider

1. Open the Google Admin console, go to `Devices` > `Chrome` > `Connectors`, and click `New provider configuration`.
2. In the list of providers, click `Configure` under `CrowdStrike XDR`.
3. Enter the API key and hostname provided by Stoïk.
4. Under `XDR events`, select `Enable XDR events`.
5. Click `Add configuration`.
6. Under `Organizational units`, find your unit and click its name.
7. Check the `CrowdStrike XDR - My XDR` configuration box.
8. Click `Save`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F6FNZr2OWMedfCxj3iDAz%2Fgoogle-admin-console-api-key-hostname-59a38cef44f6e3db8b47701f8b53635f.webp?alt=media&amp;token=e69631c3-ec1e-4156-bae5-1c2241bcd5b1" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

## Enable XDR reporting

1. Go to `Devices` > `Chrome` > `Settings`.
2. Click the `Device settings` tab.
3. In `User and device reporting`, select `Report extended detection and response (XDR) events`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FR6fgh8Kazx4PBE9ylQGd%2Fgoogle-admin-console-user-and-device-reporting-55fc96ef1aaa52f34ff5d95a4d6bc20e.webp?alt=media&amp;token=ee56795f-3430-4016-b471-6ca68c954dd7" alt=""><figcaption></figcaption></figure>
{% endstep %}
{% endstepper %}


# Linux

Install the CrowdStrike EDR on Linux endpoints with the RPM, DEB or SUSE package. For a large-scale deployment, we recommend starting with a manual test installation on a few endpoints.

{% stepper %}
{% step %}

## Download the installer

The package is only available from Stoïk Protect: go to `MDR` > `Settings`, click `Linux`, then click `Download the executable`.
{% endstep %}

{% step %}

## Find your customer ID

Go to `MDR` > `Settings` and copy your customer ID.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FUFFCuuQGmvnkhwDCjtCX%2Fimage.png?alt=media&amp;token=880090b5-c3f0-4b45-9f75-fd4b6cc0bc4c" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

## Install the package

{% tabs %}
{% tab title="RPM" %}
Red Hat Enterprise Linux, CentOS, Amazon Linux, Oracle, RHEL:

```bash
sudo yum install <package_name>
```

{% endtab %}

{% tab title="DEB" %}
Ubuntu, Debian:

```bash
sudo dpkg -i <package_name>
```

{% endtab %}

{% tab title="SUSE" %}
SLES:

```bash
sudo zypper install <package_name>
```

{% endtab %}
{% endtabs %}
{% endstep %}

{% step %}

## Activate and start the sensor

1. Activate the sensor, replacing `<customer_ID>` with your customer ID:

   ```bash
   sudo /opt/CrowdStrike/falconctl -s --cid=<customer_ID>
   ```
2. Start the sensor:

   ```bash
   systemctl start falcon-sensor
   ```

{% endstep %}
{% endstepper %}


# Virtual environments

Install the Falcon sensor on a golden image so that every cloned VM gets its own identity. The method depends on whether your VMs are **persistent** or **non-persistent**.

## Prerequisites

* A **VMware vSphere + Citrix Cloud (DaaS) MCS** environment. For other hypervisors or provisioning technologies, refer to their respective documentation. For the full Citrix DaaS documentation, see [Citrix DaaS](https://docs.citrix.com/en-us/citrix-daas).
* For non-persistent VMs, host machines that meet all three of the following criteria:
  * Are non-persistent
  * Are domain-joined
  * Use an FQDN that does not change

## VM types

Before installing, identify which type of VDI environment you are working with:

| Type                    | Description                                                                               |
| ----------------------- | ----------------------------------------------------------------------------------------- |
| **Non-persistent**      | VM does not retain changes after reboot. Rebuilt from master image on every login/reboot. |
| **Persistent (Static)** | VM retains all changes after reboot. User always connects to the same machine.            |

{% hint style="warning" %}
Using the wrong method may result in all cloned VMs sharing the same Agent ID (AID), causing detections from multiple machines to appear as coming from a single host.
{% endhint %}

## Install the sensor

{% stepper %}
{% step %}

### Prepare the golden image

The golden image (also called master image, base VM, or base image) must be properly configured before installing the Falcon sensor. The sensor must always be **the last software installed**.

#### Operating system and network

* Install Windows OS and apply all latest Windows Updates.
* Configure network settings (DNS, Gateway) and open the necessary network traffic: see [Network requirements](https://docs.stoik.io/mdr/deploy/crowdstrike/network).
* Join the machine to the Active Directory domain.
* Verify the domain trust relationship:

```powershell
Test-ComputerSecureChannel
```

Expected result: `True`

{% hint style="info" %}
Do not disable the **Windows Security Center** service. CrowdStrike requires it to be active.
{% endhint %}

#### Hypervisor tools

Install the integration tools for your hypervisor (Citrix VM Tools, Hyper-V Integration Services, or VMware Tools) **before** installing the VDA and the Falcon sensor.

#### Citrix Virtual Delivery Agent (VDA)

The VDA must be installed on the master image before creating the Machine Catalog.

1. Download the VDA from `Citrix Cloud` > `Downloads` > `Virtual Delivery Agent for Windows Single-session OS`.
2. Run the installer as administrator.
3. On the `Environment` page, select `Create a master MCS image`.
4. On the `Delivery Controller` page, enter the FQDN of your Citrix Cloud Connector: `Citrix Cloud` > `Resource Locations` > select your location > `Cloud Connectors`.
5. Complete the installation and reboot.

{% hint style="info" %}
Do not run Sysprep on master images when using MCS. MCS handles generalization and unique identity assignment for each cloned VM.
{% endhint %}
{% endstep %}

{% step %}

### Download the Falcon sensor installer

1. Log in to the Falcon Console.
2. Go to `Host Setup and Management` > `Deploy` > `Sensor Downloads`, or from Stoïk Protect: `MDR` > `Settings` > `Virtual environments` > `Download the executable`.
3. Copy your CCID (Customer ID Checksum) displayed on the same page.
4. Copy the installer onto the golden image.
   {% endstep %}

{% step %}

### Send information to Stoïk

Send the following information to <protect@stoik.io>:

* The **naming convention** of your VDI machines (e.g. `SRV-VDA-*`, `VDA-USER-###`).
* The **name of the golden image** used as the master image.

This information is required to configure the host group and the sensor update policy on Stoïk's side.
{% endstep %}

{% step %}

### Install the sensor on the golden image

{% tabs %}
{% tab title="Non-persistent" %}

#### Install

Open `cmd` as administrator on the golden image and run:

```cmd
WindowsSensor.exe /install CID=<your_CCID> VDI=1
```

#### Verify the service is running

```powershell
Get-Service CSFalconService
```

Expected: `Status = Running`

#### Leave the VM online (1 hour minimum)

After installation, leave the golden image powered on long enough for:

* The sensor to connect to the CrowdStrike cloud and update to the version defined in the assigned sensor update policy.
* All necessary files to be downloaded.

#### Shut down and take a snapshot

1. Shut down from within Windows (not from the hypervisor power off button).
2. Wait for the VM to reach **Powered Off** state in your hypervisor.
3. Take a snapshot of the powered-off VM:
   * Name: `CS-VDI-Clean-[date]`
   * Do not check `Snapshot the virtual machine's memory`.
     {% endtab %}

{% tab title="Persistent" %}
With `NO_START=1`, the sensor is installed without starting and without contacting the CrowdStrike cloud. The unique AID is assigned the first time each cloned VM boots.

{% hint style="warning" %}
After installing with `NO_START=1`, **never reboot** the golden image before shutting it down. A reboot triggers cloud communication and generates an AID that will be shared by all clones.
{% endhint %}

#### Install

Open `cmd` as administrator on the golden image and run:

```cmd
WindowsSensor.exe /install CID=<your_CCID> NO_START=1
```

#### Verify the service is stopped (expected behavior)

```powershell
Get-Service CSFalconService
```

Expected: `Status = Stopped`

The service will start automatically on the first boot of each cloned VM.

#### Shut down and take a snapshot

1. Shut down from within Windows (not from the hypervisor power off button).
2. Wait for the VM to reach **Powered Off** state in your hypervisor.
3. Take a snapshot of the powered-off VM:
   * Name: `CS-Template-Clean-[date]`
   * Do not check `Snapshot the virtual machine's memory`.
     {% endtab %}
     {% endtabs %}
     {% endstep %}

{% step %}

### Create the Machine Catalog in Citrix Cloud Web Studio

Once the golden image snapshot is ready, create or update the Machine Catalog in Citrix Cloud Web Studio. Key settings to configure in the catalog creation wizard:

| Page               | Non-Persistent                                | Persistent                                         |
| ------------------ | --------------------------------------------- | -------------------------------------------------- |
| Machine Type       | Single-session OS                             | Single-session OS                                  |
| Machine Management | MCS                                           | MCS                                                |
| Desktop Experience | **Random**                                    | **Static**                                         |
| Image              | Golden image + snapshot `CS-VDI-Clean-[date]` | Golden image + snapshot `CS-Template-Clean-[date]` |
| {% endstep %}      |                                               |                                                    |
| {% endstepper %}   |                                               |                                                    |


# Installation errors

Fix the most common errors that can occur while installing or uninstalling CrowdStrike Falcon Sensor.

## Best practices

If you encounter a block or error during installation or uninstallation:

* Never manually interrupt the installation or uninstallation process (do not force-quit in Task Manager).
* Temporarily change the computer's DNS server by modifying the network configuration.
* Temporarily disable the Windows Firewall and restart the process.
* Reinstall the Bluetooth drivers from the computer manufacturer's website.
* Before installing CrowdStrike, close/uninstall any antivirus programs already on the computer to avoid potential conflicts. See [Antivirus and exclusions](https://docs.stoik.io/mdr/deploy/antivirus).

## Troubleshooting

<details>

<summary>Network problem</summary>

Error message: *Falcon could not connect to the CrowdStrike cloud. Please check your network configuration and try again.*

In many cases, access is blocked by a central network security solution (e.g., a proxy or DPI filter). Open the necessary network connections: see [Network requirements](https://docs.stoik.io/mdr/deploy/crowdstrike/network).

</details>

<details>

<summary>Conflict with existing antivirus</summary>

Error message: *This program is not designed to run on Windows...*

This usually indicates a conflict with an already installed antivirus program. Before installing the EDR, make sure the incompatible antivirus software is uninstalled. See [Antivirus and exclusions](https://docs.stoik.io/mdr/deploy/antivirus).

</details>

<details>

<summary>Outdated Windows version</summary>

Error message: *CrowdStrike Falcon for Windows x64 is only compatible with Windows 10, Windows 11...*

The installer you are using is generally not compatible with your Windows version. Download the correct version of the installer for your operating system.

</details>

<details>

<summary>CrowdStrike already installed</summary>

Error message: *Enter the Falcon maintenance token for this host.*

The executable file is already installed on your computer.

</details>


# Add-on modules

Set up optional CrowdStrike modules: mobile devices, Device Control and ITP with Entra ID. To buy an additional option, contact <protect@stoik.io>.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-mobile">:mobile:</i></h4></td><td><h4>Mobile devices</h4></td><td>Protect iOS and Android phones, through an MDM or manually.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/add-on/mobile">Mobile devices</a></td></tr><tr><td><h4><i class="fa-usb">:usb:</i></h4></td><td><h4>Device Control</h4></td><td>Control which USB devices can be used on your endpoints.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/add-on/device-control">Device Control</a></td></tr><tr><td><h4><i class="fa-microsoft">:microsoft:</i></h4></td><td><h4>ITP with Entra ID</h4></td><td>Detect identity threats on your Microsoft Entra ID accounts.</td><td><a href="https://docs.stoik.io/mdr/deploy/crowdstrike/add-on/itp">ITP with Entra ID</a></td></tr></tbody></table>


# Mobile devices

Install CrowdStrike Falcon on smartphones, either through an MDM solution or manually with a QR code.

## Installation

{% tabs %}
{% tab title="MDM" %}
CrowdStrike Falcon installs through an MDM like any other managed application: push it from your MDM console (app store or enterprise catalog) to the target devices, making sure the customer ID is configured for the app.

There are too many MDM solutions to document each one individually. Refer to your MDM provider's own documentation for the exact deployment steps.

{% hint style="warning" %}
Deploying via MDM does not remove the manual step described in the `Manual` tab: each employee still needs to receive and scan their QR code by email to complete the setup.
{% endhint %}
{% endtab %}

{% tab title="Manual" %}

#### Setup in Stoïk Protect

1. In Stoïk Protect, go to `MDR` > `Settings`.
2. In the `Mobiles` section, click **Deploy**.
3. Add employees by entering their email addresses manually, or uploading a CSV file.
4. Each employee receives a personalized email containing a QR code.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-1eb0fe742b352753f0af4712fbdb732f606669a8%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2015.53.14%20(1).png?alt=media" alt=""><figcaption></figcaption></figure>

#### Install on the mobile device

1. Download the CrowdStrike Falcon app.
2. Open the app and scan the QR code received by email.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fp3Y7wn3auaeILwmNkpSX%2Fimage%20(13)-2.webp?alt=media&amp;token=562efe02-97ba-4357-8cad-94a5412787c4" alt=""><figcaption></figcaption></figure>

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FruDRrtjnGGKQp6s4j8o5%2Fimage%20(14)-1.webp?alt=media&amp;token=701d4845-6ee4-4707-8f88-9164b5df7dd5" alt=""><figcaption></figcaption></figure>

The EDR installs automatically and the device appears in your monitored hosts list in Stoïk Protect.
{% endtab %}
{% endtabs %}

## Personal use of work phones

* **Android:** Two separate environments can be configured via software separation (with two "Play Stores"). In this case, no personal activity is blocked, as CrowdStrike is only installed on one environment.
* **iOS:** CrowdStrike does not have the ability to block applications. The installation is non-intrusive, focused on detecting risky behavior.

## What the mobile EDR blocks

* Malicious IP addresses and domains
* Applications deemed dangerous

The features for viewing installed applications are not enabled.


# Device Control

Control the use of USB devices (USB keys, hard drives, etc.) with a USB device **Policy**: general rules that allow or block USB devices based on their class. The Stoïk team can set it up for you, or you can configure it yourself in the CrowdStrike console.

## Prerequisites

* For self-service setup: **Device Control Manager** access to your CrowdStrike console. Request it by emailing <protect@stoik.io>.

## Scope

You can:

* Configure rules by device type.
* Implement them using CrowdStrike Policies, either yourself or through the Stoïk SOC.
* Configure exclusions (workstation, device).
* Manage them with the `Device Control Manager` user right.

Configuration by endpoint is not possible: it would require assigning each endpoint to specific **groups**.

For example, you can create a policy to block USB storage devices while allowing access to other USB device classes.

## Deployment

{% tabs %}
{% tab title="Stoïk team" %}

#### Define your control objectives

Determine what you want to enforce:

* Block all USB devices
* Set devices to read-only
* Allow only specific models
* Monitor / audit usage without blocking (observation mode)

#### Inventory of authorized or blocked devices

For each USB device you want to allow or block:

1. Plug it into a workstation.
2. Retrieve its hardware identifiers, either:
   * Via the graphical interface: `Device Manager` > USB device > right-click > `Properties` > `Details` > `Hardware IDs`.
   * Via PowerShell:

     ```powershell
     Get-PnpDevice -PresentOnly | Where-Object { $_.Class -eq "USB" }
     ```
3. Send the following information to <protect@stoik.io>:
   * Vendor ID (VID)
   * Product ID (PID)
   * Serial number (if applicable)

#### Stoïk team configuration

A Stoïk cybersecurity engineer configures the rules and informs you once the setup is complete.
{% endtab %}

{% tab title="Self-service" %}
This option requires **Device Control Manager** access, see [Prerequisites](#prerequisites).

#### Create and configure the policy

1. Go to `Endpoint Security` > `USB Device Control` > [`Policies`](https://falcon.eu-1.crowdstrike.com/policies/device-control).
2. Click `Create policy`, enter a name and optional description, then click `Create policy` to confirm.
3. In the `Settings` tab:
   * Select the mode (`Monitor`, `Monitor and enforce`, or `Off` for macOS only).
   * Enable the desired USB settings.
   * Select the notification to inform end users when a USB device is fully blocked.
   * Select the notification to inform end users when a USB device is restricted.
4. In the `USB Device` tab, expand each device class to include, choose the authorization level, and add exceptions per class if needed by clicking `Add exception`.

The three modes work as follows:

* **Monitor and enforce**: actively applies defined rules, blocks or allows USB connections and logs activity.
* **Monitor only**: logs connections without enforcing restrictions. Ideal for testing policies without disrupting users.
* **Off (macOS only)**: completely disables Device Control. No activity is logged or blocked.

{% hint style="info" %}
Using **Monitor only** is an effective way to establish a list of USB devices typically used by your employees, which can then serve as the basis for a future allowlist.
{% endhint %}

<figure><img src="https://25674881.fs1.hubspotusercontent-eu1.net/hubfs/25674881/Screenshot%202025-08-20%20at%2009-58-35-png.png" alt=""><figcaption></figcaption></figure>

<figure><img src="https://25674881.fs1.hubspotusercontent-eu1.net/hubfs/25674881/Screenshot%202025-08-20%20at%2009-59-34-png.png" alt=""><figcaption></figcaption></figure>

#### Activate the policy

1. Click `Save` in the top-right corner, then `Save` again to confirm.
2. In the `Assigned host groups` tab, click `Assign host groups to policy`.
3. Click `Enable policy`.

To speed up configuration of similar policies, click `Duplicate policy` to create an exact copy.
{% endtab %}
{% endtabs %}

## Policy deployment

{% tabs %}
{% tab title="Windows" %}
Device policies are applied as soon as a USB device is connected to a host. If USB devices are already connected at the time the policy is assigned, they will only be affected upon their next reconnection or the next host reboot.
{% endtab %}

{% tab title="macOS" %}
Device policies are applied as soon as a USB or Bluetooth device is connected to a host. If devices are already connected at the time the policy is assigned, the policy takes effect immediately. If the policy blocks the device, it will be disconnected.
{% endtab %}
{% endtabs %}


# ITP with Entra ID

Connect each Entra ID tenant to CrowdStrike Identity Threat Protection (ITP) by creating an Azure registration from your Falcon console. At the end, you email Stoïk your tenant ID so monitoring can be activated.

## Prerequisites

* A **Microsoft Entra ID P1 or P2** licence. Microsoft Entra ID Free is not supported.
* An **Azure subscription** in the tenant. CrowdStrike deploys its log ingestion resources there.
* A person with the **Global Administrator** or **Privileged Role Administrator** role in Entra ID, to grant admin consent.
* A person with the **Owner** and **Azure Deployment Stack Owner** Azure roles at the **Tenant Root Group** level, to deploy the resources.
* A **Falcon console account** with the permissions to set up the connector. Stoïk creates it for you in step 1.

{% hint style="warning" %}
The Global Administrator role in Entra ID does **not** grant the Azure roles above. They must be assigned explicitly at the Tenant Root Group level, otherwise the deployment fails.
{% endhint %}

## Purpose of ITP

ITP continuously monitors account activity and Active Directory to detect abnormal behaviour.

* **Identity-centred behavioural analysis:** detects suspicious authentications, unusual access patterns, unauthorised login attempts and user account anomalies. Examples include suspicious workstation-to-server connections, network scans, brute-force attempts and credential extraction. The module can automatically block critical events.
* **Active Directory security posture:** gives a real-time view of AD security, with more advanced options than the Stoïk Protect AD scan. It identifies configuration issues, accounts with compromised passwords and privilege management problems.
* **Cloud threat monitoring (optional):** once Entra ID is connected, ITP ingests your cloud and hybrid accounts, Entra ID audit logs, sign-in logs and Azure activity logs. It correlates SSO activity with your on-premises accounts and detects logins from suspicious IP addresses and abnormal access attempts.

## Integration

CrowdStrike changed this integration in 2026: the Entra ID connector is now set up as an **Azure registration** in Falcon Cloud Connectors. You no longer create an application, a client secret or a custom role by hand. If you already set up the previous connector, see [Migrating from the legacy connector](#migrating-from-the-legacy-connector).

You run the whole setup yourself from your Falcon console, for **each Entra ID tenant**. Stoïk gives you the access you need and checks the result. **No secret is exchanged with Stoïk**: you generate the CrowdStrike API credentials in the console and only use them in your own Azure Cloud Shell.

Ideally, the same person runs steps 3 to 6 and holds both the Entra ID and Azure roles above. Otherwise, plan a short session with the people concerned: the wizard must stay open from step 3 to step 6.

<details>

<summary>Who does what</summary>

<table data-search="false"><thead><tr><th>Step</th><th width="320">Who</th></tr></thead><tbody><tr><td>1. Get your Falcon console access</td><td>You and Stoïk</td></tr><tr><td>2. Check your Azure roles</td><td>You (Azure administrator)</td></tr><tr><td>3. Create the Azure registration in Falcon</td><td>You</td></tr><tr><td>4. Grant admin consent</td><td>You (Global Administrator or Privileged Role Administrator)</td></tr><tr><td>5. Deploy the resources from Azure Cloud Shell</td><td>You (Azure administrator)</td></tr><tr><td>6. Complete and confirm the registration</td><td>You, then Stoïk</td></tr><tr><td>7. Grant remediation access to Stoïk</td><td>You (Global Administrator)</td></tr></tbody></table>

</details>

{% stepper %}
{% step %}

### Get your Falcon console access

1. Email <protect@stoik.io> with the **name and email** of the person who will run the setup.
2. Stoïk creates a Falcon user account for this person, with the permissions needed to configure the Entra ID connector.
3. The person receives an activation email from CrowdStrike. They set their password and **MFA**, then sign in to the [Falcon console](https://falcon.eu-1.crowdstrike.com).
   {% endstep %}

{% step %}

### Check your Azure roles

The person running the deployment opens [Azure Cloud Shell](https://shell.azure.com) in **Bash** mode and runs:

```bash
az role assignment list \
  --assignee $(az ad signed-in-user show --query id -o tsv) \
  --scope "/providers/Microsoft.Management/managementGroups/$(az account show --query tenantId -o tsv)" \
  --include-groups \
  --query "[?roleDefinitionName=='Owner' || roleDefinitionName=='Azure Deployment Stack Owner'].{Role:roleDefinitionName, Scope:scope}" \
  -o table
```

The output must list both **Owner** and **Azure Deployment Stack Owner**. If either role is missing, have it assigned at the Tenant Root Group level before you continue.

<details>

<summary>Can't see or manage the Tenant Root Group</summary>

A Global Administrator must first enable `Access management for Azure resources` in `Entra ID` > `Properties`. This gives them the right to assign the roles above. You can turn it off again afterwards.

</details>
{% endstep %}

{% step %}

### Create the Azure registration in Falcon

Before you start, copy from your Microsoft consoles:

* the `Tenant ID`: [Microsoft Entra admin center](https://entra.microsoft.com) > `Entra ID` > `Overview`
* the `Subscription ID` where CrowdStrike resources will be deployed: [Azure portal](https://portal.azure.com) > `Subscriptions`

Then, in the Falcon console:

1. **Open the wizard:**
   * If you have a legacy Entra ID connector, go to `Identity protection` > `Settings` > `Configurations` > `Connectors` and click **Upgrade Now** on it. The wizard opens with your tenant pre-filled.
   * Otherwise, go to `Connectors` > `Cloud connectors` > `Cloud connections`, open the `Azure` tab and click **Add new registration**. If prompted, select **New deployment methods**, then **Next**.
2. **Registration parameters:** enter a registration name, your `Microsoft Entra tenant ID`, select `Bicep` and enter the `Subscription ID`. Click **Continue with standard setup**.
3. **Features:** select `Real-time visibility and detection`, then **Next**.
4. **Advanced settings:** leave `I want to customize Entra ID permissions` unchecked. Change the region for log ingestion resources only if needed. Click **Next**.
5. **API credentials:** click **Generate API client**. Copy the `Client ID` and `Client secret` and keep them at hand for step 5. The secret is shown only once. Click **Done**, then **Next**.

{% hint style="warning" %}
Keep the client secret to yourself. Never send it by email, chat or ticket, including to Stoïk. You only need it in Azure Cloud Shell during step 5.
{% endhint %}
{% endstep %}

{% step %}

### Grant admin consent

Run by a Global Administrator or Privileged Role Administrator. Without admin consent, Falcon cannot read Entra ID data and Identity Protection will not work.

1. On the `Deploy configuration` page of the wizard, click **Grant admin consent** and sign in with your Microsoft administrator account.
2. Check that the **tenant ID in the URL matches your tenant**.
3. Review the requested permissions and click **Accept**.
4. Back in the wizard, click **Verify**.

The application created by CrowdStrike only has **read-only** permissions on your directory.
{% endstep %}

{% step %}

### Deploy the resources from Azure Cloud Shell

1. On the `Deploy configuration` page, open the `Bash` tab and click **Download ZIP** to get `crowdstrike-fcs-azure-bicep-generated.zip`.
2. In Azure Cloud Shell (Bash), upload the ZIP file.
3. Unzip it:

   ```bash
   unzip crowdstrike-fcs-azure-bicep-generated.zip
   ```
4. Check that the parameters file is present:

   ```bash
   ls parameters.bicepparam
   ```

   If the file isn't listed, check that you uploaded and unzipped the right file.
5. Copy the commands shown on the `Bash` tab and run them in order:
   * `az login -t '<tenant ID>'`
   * `export FALCON_CLIENT_SECRET=<client secret from step 3>`
   * `az stack mg create ...`, which deploys the Bicep template
6. Wait until the last command finishes. The terminal looks inactive while the deployment runs and only shows errors. This is normal.
   {% endstep %}

{% step %}

### Complete and confirm the registration

1. In the wizard, click **Validate and complete**.
2. On the `Azure` tab of `Cloud connections`, the registration first shows `Pending`, then `Active`. Allow about **2 hours** for data to appear.
3. The registration is complete when both `Assets` and `Real-time visibility and detection` show `Active`.

{% hint style="warning" %}
Email <protect@stoik.io> with your tenant ID. Stoïk checks the registration and activates monitoring on its side. Without it, monitoring cannot be activated.
{% endhint %}

Once the registration is active, Stoïk removes the configuration permissions from your Falcon account.
{% endstep %}

{% step %}

### Grant remediation access to Stoïk

If an account is compromised outside business hours, your team may be unavailable. Granting this access lets Stoïk disable affected accounts once a compromise is confirmed.

1. Sign in as a Global Administrator and open the [admin consent link](https://login.microsoftonline.com/organizations/adminconsent?client_id=72f25200-dfe6-4c66-959d-103731ae6215\&redirect_uri=https://login.microsoftonline.com/common/oauth2/nativeclient).
2. Review the requested permissions and click **Accept**.
3. You land on a blank page. You can close it.

<details>

<summary>What the remediation application can do</summary>

The application can only enable or disable existing accounts. It cannot create accounts or change other attributes. Each action requires MFA, is logged and is timestamped. Stoïk can provide the full audit log on request.

</details>
{% endstep %}
{% endstepper %}

## Migrating from the legacy connector

If you set up the [previous Entra ID connector](https://docs.stoik.io/mdr/deploy/crowdstrike/add-on/itp-legacy) (application `stoik-connector-itp` with a client secret), it must be migrated:

1. Request your Falcon console access (step 1).
2. Complete steps 2 to 6. In step 3, start from **Upgrade Now** on the legacy connector.
3. Once the new registration is active, the legacy connector shows `Ready to Decommission`. Delete it in `Identity protection` > `Settings` > `Configurations` > `Connectors`, or ask Stoïk to do it.
4. You can then delete the old `stoik-connector-itp` application, its client secret and the `Identity Protection RBAC Reader` custom role from your tenant. They are no longer used.

## Self-managed Falcon console

If you manage your own Falcon console, skip step 1 and use an account with the permissions to manage cloud registrations and API clients. Follow steps 2 to 7.


# ITP with Entra ID (legacy)

Set up the legacy Entra ID connector for CrowdStrike ITP by creating an application in Entra ID. This connector is **deprecated** and only available in specific cases, on request to Stoïk: use the [current setup](https://docs.stoik.io/mdr/deploy/crowdstrike/add-on/itp) instead.

## Prerequisites

* A paid Entra ID license. [Microsoft Entra ID](https://entra.microsoft.com) Free is not supported.
* An account with global administrator rights to your Entra ID account.

## Purpose of ITP

The ITP (Identity Threat Protection) option continuously monitors account-related activities and Active Directory to detect abnormal behavior.

### Identity-centered behavioral analysis

* Suspicious authentications
* Unusual access patterns
* Unauthorized login attempts
* User account anomalies

Examples of detected events: suspicious workstation-to-server connections, network scans, brute-force attempts, or credential extraction. When a critical event is identified, the module can **automatically block it**.

### Assessment of your Active Directory security posture

The module provides a real-time overview of your Active Directory security, similar to the Stoïk Protect AD scan but with more advanced options:

* Identification of configuration issues
* Detection of accounts with compromised passwords
* Management and control of user privileges

### Cloud threat monitoring (optional)

When configured with your identity service (e.g. Entra ID), ITP can detect threats in your cloud environments:

* Logins from suspicious IP addresses
* Abnormal access attempts to user accounts

## Integration

The integration requires the following steps for **each Azure tenant**.

{% stepper %}
{% step %}

### Create and configure an application in Entra ID

1. In the Azure portal, open the left-hand menu and click `Microsoft Entra ID`.
2. From the `Add` menu, select `App registrations`.
3. On the `Register an application` page:
   * Under `Name`, enter a name matching your internal conventions, or use `stoik-connector-itp`.
   * Under `Supported account types`, select `Accounts in this organizational directory only`.
   * Under `Redirect URI`, select `Web` and enter `https://localhost`.
   * Click `Register`.
4. In the application's left-hand menu, click `API permissions`.
5. Click `Add a permission`.
6. In the `Request API permissions` panel, click `Microsoft Graph`, then select `Application permissions`.
7. In `Select permissions`:
   * Type `directory`, expand `Directory`, and select `Directory.Read.All`.
   * Type `audit`, expand `AuditLog`, and select `AuditLog.Read.All`.
8. Click `Add permissions`, then click `Grant admin consent` and confirm.
   {% endstep %}

{% step %}

### Retrieve the tenant domain

1. In the Azure portal, go to `Microsoft Entra ID`.
2. In the left-hand menu, click `Overview`.
3. Note the value of the `Primary domain` (e.g. `example.onmicrosoft.com`).
   {% endstep %}

{% step %}

### Retrieve the application ID and secret

1. In Microsoft Entra ID, go to `App registrations` and select your application.
2. In `Overview`, note the `Application (client) ID`.
3. In `Certificates & secrets`, create a new client secret: provide a description and an expiration duration.
4. Save the generated value immediately. It will not be visible again.

<figure><img src="https://25674881.fs1.hubspotusercontent-eu1.net/hubfs/25674881/Entra_value-71a566ca11435df656075b4c186be741-png-1.png" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Create a custom role for RBAC access

To display Azure RBAC role assignment information in Identity Protection, the application must be granted permission to read RBAC definitions and assignments.

#### Create the custom role

1. In the Azure portal, go to `Management Groups`.
2. Click `Tenant Root Group`.
3. Select `Access control (IAM)`, then click `Roles`.
4. Click `Add`, then select `Add custom role`.
5. In the `Basics` tab, enter a clear name (e.g. `Identity Protection RBAC Reader`).
6. In the `Permissions` tab, add the following permissions:

   ```
   Microsoft.Authorization/roleDefinitions/read
   ```

   ```
   Microsoft.Authorization/roleAssignments/read
   ```
7. Click `Review + create`, then `Create`.

#### Assign the custom role

1. In the Azure portal, go to `Management Groups`.
2. Click `Tenant Root Group`.
3. Select `Access control (IAM)`, then click `Role assignments`.
4. Click `Add`, then select `Add role assignment`.
5. In the `Role` tab, select the custom role you just created.
6. In the `Members` tab, click `Select members`, search for the previously created application (e.g. `stoik-connector-itp`), select it, and click `Select`.
7. Click `Review + assign`, then `Review + assign` again to confirm.
   {% endstep %}

{% step %}

### Transmit information to Stoïk

{% hint style="info" %}
If you have direct access to your Falcon ITP platform, skip this step and proceed to the next one.
{% endhint %}

Send the following information to <protect@stoik.io>:

* Tenant primary domain name (ends with `onmicrosoft.com`)
* Application ID
* Application secret

Without this information, the connector cannot be activated on our side.
{% endstep %}

{% step %}

### Configure the connector in Identity Protection

1. In the Falcon console, go to `Identity Protection` > `Configure` > `Connectors`.
2. Select `Entra` under the `IDAAS` category.
3. Configure the connector with the tenant primary domain name, the application ID, and the application secret.
4. Enable the login data collection option.
5. Save the configuration.

<figure><img src="https://25674881.fs1.hubspotusercontent-eu1.net/hubfs/25674881/image-png-Jan-28-2025-10-22-05-6569-AM.png" alt=""><figcaption></figcaption></figure>

The status indicator turns green within a minute, confirming the connection has been successfully established.

{% hint style="info" %}
The tenant primary domain name to use is the one generated by Microsoft during the creation of your Azure tenant. It always ends with `onmicrosoft.com`.
{% endhint %}
{% endstep %}

{% step %}

### Grant remediation access to Stoïk

If an account is compromised during nights or weekends, your teams may not be available to perform remediation. We recommend granting access so impacted accounts can be disabled once a compromise is confirmed.

Click the [provided link](https://login.microsoftonline.com/organizations/adminconsent?client_id=72f25200-dfe6-4c66-959d-103731ae6215\&redirect_uri=https://login.microsoftonline.com/common/oauth2/nativeclient), review the requested permissions, and click `Accept`. You will be redirected to a blank page, which you may close.

<figure><img src="https://25674881.fs1.hubspotusercontent-eu1.net/hubfs/25674881/image-png-Feb-20-2025-09-10-33-4687-PM.png" alt=""><figcaption></figcaption></figure>

The application only has the permissions required to enable or disable existing accounts, with no ability to create accounts or modify other attributes. Each action is authenticated via MFA, logged, and timestamped. A complete audit log can be provided upon request.
{% endstep %}
{% endstepper %}


# Network requirements

If outbound traffic is blocked, open port **443** for the following addresses. These apply to every operating system, not only Windows.

No inbound ports are required. All communication is outbound to CrowdStrike.

## IP addresses, domains and DNS names

### IPv4

* 3.121.6.180
* 3.121.187.176
* 3.121.238.86
* 3.125.15.130
* 18.158.187.80
* 18.198.53.88
* 3.78.32.129
* 3.121.13.180
* 3.123.240.202
* 18.184.114.155
* 18.194.8.224
* 35.156.219.65
* 3.69.184.79
* 3.76.143.53
* 3.77.82.22
* 18.197.35.253
* 3.122.135.178
* 18.194.222.197

### IPv6

* `2a05:d014:45e:4e00::/56`

### Domains

* `ts01-lanner-lion.cloudsink.net`
* `lfoup01-lanner-lion.cloudsink.net`
* `lfodown01-lanner-lion.cloudsink.net`

### Public DNS

* `https://falcon.eu-1.crowdstrike.com`
* `https://assets.falcon.eu-1.crowdstrike.com`
* `https://assets-public.falcon.eu-1.crowdstrike.com`


# SentinelOne

Install the SentinelOne EDR on your workstations and servers, then review the add-on modules and network requirements. Once the EDR is set up, monitoring starts automatically and full protection is active as soon as it is installed. No action is required on your part unless otherwise requested.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-laptop">:laptop:</i></h4></td><td><h4>Operating systems</h4></td><td>Install the EDR on Windows, macOS or Linux.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/os">Operating systems</a></td></tr><tr><td><h4><i class="fa-puzzle-piece">:puzzle-piece:</i></h4></td><td><h4>Add-on modules</h4></td><td>Set up optional modules such as Identity Security.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/add-on">Add-on modules</a></td></tr><tr><td><h4><i class="fa-network-wired">:network-wired:</i></h4></td><td><h4>Network requirements</h4></td><td>Open the domains and IP addresses the EDR needs.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/network">Network requirements</a></td></tr></tbody></table>


# Operating systems

Install the SentinelOne EDR on each operating system in your environment. A simplified, guided installation walkthrough is also available in Stoïk Protect, under `MDR` > `Settings`. The guides below include screenshots.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-windows">:windows:</i></h4></td><td><h4>Windows</h4></td><td>Install the EDR manually, by GPO or through an MDM.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/os/windows">Windows</a></td></tr><tr><td><h4><i class="fa-apple">:apple:</i></h4></td><td><h4>macOS</h4></td><td>Install and activate the agent manually or through an MDM.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/os/macos">macOS</a></td></tr><tr><td><h4><i class="fa-linux">:linux:</i></h4></td><td><h4>Linux</h4></td><td>Install the package and activate the agent.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/os/linux">Linux</a></td></tr></tbody></table>


# Windows

Install the SentinelOne agent on a Windows workstation or server manually, through an MDM or with a GPO. To deploy the EDR on a large scale, start with a manual test installation on a few endpoints.

## Prerequisites

* Your current antivirus is completely uninstalled: SentinelOne replaces it. If the firewall is managed separately, keep it running. If it is part of the same solution, copy any custom rules to Windows Firewall first.
* The machine has been restarted after removing the previous antivirus and after the last Windows updates.
* You are logged in with an administrator account on the machine.
* If outbound traffic is blocked, the addresses in [Network requirements](https://docs.stoik.io/mdr/deploy/sentinelone/network) are allowed.

## Installation

{% stepper %}
{% step %}

### Download the right installer

The installer depends on the Windows version and the processor of the machine. To find them, press the Windows key, type `About your PC` and press Enter. Read the `Edition` (Windows version) and `System type` (processor) lines. On Windows 7 or 8, open the `Start` menu, right-click `Computer` and select `Properties` instead.

In Stoïk Protect, go to `MDR` > `Settings` and click `Windows`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-b76c8b283ffc7b58936e5f3915185f1c3af8df6e%2Fs1-windows-setup.webp?alt=media" alt=""><figcaption></figcaption></figure>

Download the file that matches the machine:

| The machine has                                                       | Menu in Stoïk Protect      | File to download                                           |
| --------------------------------------------------------------------- | -------------------------- | ---------------------------------------------------------- |
| Windows 8.1 or later, 64-bit, x64 processor (Intel / AMD)             | `Download this executable` | `SentinelInstaller_windows_64bit_v<version>.msi`           |
| Windows 8.1 or later, ARM processor                                   | `Download this executable` | `SentinelOneInstaller_windows_arm64_v<version>.exe`        |
| Windows older than 8.1 (7, 8, XP, Vista, Server 2003 to 2012), 64-bit | `Legacy Plus installer`    | `SentinelOneInstaller-64bit_windows_legacy_v<version>.exe` |
| Any Windows, 32-bit                                                   | `Legacy Plus installer`    | `SentinelOneInstaller-32bit_windows_legacy_v<version>.exe` |

<details>

<summary>Supported Windows versions</summary>

**Standard installer** (`Download this executable` menu):

* Windows 8.1 / 10 / 11 / 11 23H2 / 11 24H2 64-bit
* Windows Server IoT 2019 / 2022 / 2025
* Windows Server / Server Core 2012 R2 / 2016 / 2019 / 2022 / 2025

**Legacy Plus installer** (`Legacy Plus installer` menu, 32-bit and 64-bit files):

* Windows 8.1 / 10 32-bit
* Windows 7 SP1 / 8 32/64-bit
* POSReady 7
* Windows Server / Storage Server / Server Core 2012 (not R2) 32/64-bit
* Windows Server 2008 R2 SP1 32/64-bit
* Windows XP SP3 / Vista SP2 32/64-bit
* Windows XP SP2 64-bit (AMD64/EM64T)
* Windows Embedded POSReady 2009
* Server 2003 SP2 / 2003 R2 SP2 / 2008 SP2 (not R2) 32/64-bit

</details>

{% hint style="warning" %}
Most machines have an Intel or AMD (x64) processor. For them, download both the `.msi` (to install with a double-click) and the x64 `.exe` (`SentinelOneInstaller_windows_64bit_v<version>.exe`, to install from the command prompt if the `.msi` fails). ARM and Legacy Plus versions are rare.

Save the file on the machine itself, for example on the Desktop. Do not run it from a USB key or a network share.
{% endhint %}
{% endstep %}

{% step %}

### Find your site token

Go to `MDR` > `Settings` and copy the site token, shown as `Customer ID`. The installer asks for it in the `Site Token` field.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FemeExV0LCLaQAfG1qQgL%2Fimage.png?alt=media&amp;token=99f2d142-fd0d-4753-9c72-f1002e271f6d" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

### Install SentinelOne

{% tabs %}
{% tab title="Manual" %}
For an Intel / AMD (x64) machine, with the `.msi`:

1. Double-click the `.msi` file.
2. Accept the license agreement.
3. Paste your site token in the `Site Token` field.
4. Click `Install`. When Windows asks `Do you want to allow this app to make changes to your device?`, click `Yes` (with a standard account, an administrator types their name and password).
5. Wait for the end of the installation, then click `Finish`.

{% hint style="warning" %}
Never double-click the x64 `.exe`: it does nothing and no longer asks for the site token. Use the `.msi` for a double-click installation, and the `.exe` only from the command prompt (see [Troubleshooting](https://docs.stoik.io/mdr/deploy/sentinelone/os/windows/troubleshooting)).
{% endhint %}

<details>

<summary>Legacy Plus machine (older than 8.1, or 32-bit)</summary>

1. Double-click the Legacy Plus file.
2. If Windows asks `Do you want to allow this app to make changes to your device?`, click `Yes`, or have an administrator type their name and password.
3. Paste your site token in the `Site Token` field, then click `Install`.

The command prompt methods do not apply to the Legacy Plus installer. If it fails, go through [Troubleshooting](https://docs.stoik.io/mdr/deploy/sentinelone/os/windows/troubleshooting), then contact <protect@stoik.io>.

</details>

<details>

<summary>ARM machine</summary>

The ARM installer has no installation window: it only works from the command prompt.

1. In `C:\`, create a folder `Temp` and, inside it, a folder `S1`. Move the installer into `C:\Temp\S1`.
2. Press the Windows key, type `cmd`, right-click `Command Prompt`, select `Run as administrator` and click `Yes`.
3. Type the commands below, pressing Enter after each line. Replace `<version>` with the version in your file name and `<site_token>` with your site token, and remove the `<` and `>` characters.

```
cd C:\Temp\S1
SentinelOneInstaller_windows_arm64_v<version>.exe -t "<site_token>" -q
```

4. Wait about 5 minutes, then type `echo %errorlevel%` and press Enter. `0` means the installation succeeded. Any other number: see [Troubleshooting](https://docs.stoik.io/mdr/deploy/sentinelone/os/windows/troubleshooting).

</details>
{% endtab %}

{% tab title="MDM" %}
Run the following command, replacing `<customer_ID>` with your customer ID:

```
<file_name>.exe -t <customer_ID> -q
```

{% endtab %}

{% tab title="GPO" %}

1. Place the `.exe` installer in a network share. Example: `\\Server\Share\SentinelOneInstaller_windows_64bit_v<version>.exe`. Give **Read** permissions on the share to `Domain Computers` (the computer accounts), not only to users: the script runs as the machine, not as a user.
2. Create a deployment script named `InstallSentinelOne.bat` with the content below. It is a plain batch script: it runs in the command prompt, which is what the SentinelOne `.exe` requires.

```batch
@echo off
set "INSTALLER=\\<server>\<share>\SentinelOneInstaller_windows_64bit_v<version>.exe"
set "SITETOKEN=<customer_ID>"
set "LOG=C:\Windows\Temp\S1_GPO_Install.log"

:: Skip if SentinelOne is already installed
sc query SentinelAgent >nul 2>&1
if %errorlevel% equ 0 (
    echo %date% %time% SentinelOne already installed >> "%LOG%"
    exit /b 0
)

echo %date% %time% Starting installation from %INSTALLER% >> "%LOG%"
start "" /wait "%INSTALLER%" -t "%SITETOKEN%" -q
echo %date% %time% Installer exit code: %errorlevel% >> "%LOG%"
```

Replace the installer path with your share path and `<customer_ID>` with your customer ID. Remove the `<` and `>` characters as well. The check at the top stops the script from reinstalling SentinelOne at every restart.

3. Create a GPO:
   1. Open the Group Policy Management Console: `Windows + R` > `gpmc.msc` > `Enter`.
   2. Right-click the target container (domain or OU) and click `Create a Group Policy Object in this domain and link it here`. Name it (e.g. `SentinelOne Deployment`).
   3. Right-click the new GPO and select `Edit`.
   4. Go to `Computer Configuration` > `Policies` > `Windows Settings` > `Scripts (Startup/Shutdown)`. Use `Computer Configuration`, not `User Configuration`: startup scripts run with system rights, so there is no administrator prompt and users have nothing to do.
   5. Double-click `Startup`, click `Show Files`, and copy `InstallSentinelOne.bat` into the folder that opens.
   6. Click `Add`, then `Browse`, select `InstallSentinelOne.bat` and click `OK`.
4. Apply the GPO:
   1. On a target machine, run `gpupdate /force`, then restart it to trigger the script.
   2. Check that the machine appears in Stoïk Protect under `MDR` > `Monitored Assets`.
   3. If it does not, open `C:\Windows\Temp\S1_GPO_Install.log` on that machine and see [Troubleshooting](https://docs.stoik.io/mdr/deploy/sentinelone/os/windows/troubleshooting).
      {% endtab %}
      {% endtabs %}
      {% endstep %}

{% step %}

### Check the installation

* On the machine, look for the SentinelOne icon in the notification area, at the bottom right of the taskbar (click `^` to see hidden icons).
* Within 5 minutes, the machine appears in Stoïk Protect under `MDR` > `Monitored Assets`. No restart is needed.

If the machine does not appear after 15 minutes, see [Troubleshooting](https://docs.stoik.io/mdr/deploy/sentinelone/os/windows/troubleshooting).
{% endstep %}
{% endstepper %}

## Troubleshooting

If the installation fails or the machine does not connect, see [Troubleshooting](https://docs.stoik.io/mdr/deploy/sentinelone/os/windows/troubleshooting): it covers how to locate the problem, administrator rights, the command prompt fallback, connection issues, and the exit codes.


# Troubleshooting

When a SentinelOne installation fails or the machine does not connect, go through the sections below in order. Each one tells you whether to move on to the next.

## Locate the problem

Find out whether the installation failed or the agent is installed but cannot reach the platform.

{% tabs %}
{% tab title="Without commands" %}

1. Press the Windows key, type `Add or remove programs` and press Enter.
2. In the search field of the list, type `Sentinel`.

| What you see                        | What it means                                                | What to do                                                                               |
| ----------------------------------- | ------------------------------------------------------------ | ---------------------------------------------------------------------------------------- |
| `Sentinel Agent` is not in the list | The installation failed                                      | Go to [Run the .msi with administrator rights](#run-the-.msi-with-administrator-rights). |
| `Sentinel Agent` is in the list     | SentinelOne is installed but cannot reach the Stoïk platform | Go to [Installed but not connected](#installed-but-not-connected).                       |
| {% endtab %}                        |                                                              |                                                                                          |

{% tab title="With commands (IT teams)" %}
Open a command prompt or PowerShell as administrator: press the Windows key, type `cmd` (or `PowerShell` for the last check), right-click it, select `Run as administrator` and click `Yes`. The window title starts with `Administrator:`.

Is the agent installed and running?

```
sc.exe query SentinelAgent
```

* `The specified service does not exist` (error 1060): the agent is not installed. Go to [Run the .msi with administrator rights](#run-the-.msi-with-administrator-rights).
* `STATE : 4 RUNNING`: the agent is installed and running. Check its connection below.
* Any other state: the agent is installed but not running. Restart the machine, then check again.

Is the agent connected to the platform?

```
cd "C:\Program Files\SentinelOne\Sentinel Agent*"
sentinelctl management status
```

If `cd` fails, type `cd "C:\Program Files\SentinelOne\` and press Tab to complete the folder name. If the agent is not connected to the management server, go to [Installed but not connected](#installed-but-not-connected).

Can the machine reach the platform? In PowerShell, this tests port 443 to each domain in [Network requirements](https://docs.stoik.io/mdr/deploy/sentinelone/network). It can take a few minutes.

```powershell
$domains = "euce1-100","euce1-102","euce1-103","euce1-104","euce1-105","euce1-106","euce1-108","euce1-109","euce1-120-mssp","euce1-ir","euwe3-801"
$domains | ForEach-Object { $h = "$_.sentinelone.net"; [pscustomobject]@{ Domain = $h; Port443 = (Test-NetConnection $h -Port 443 -WarningAction SilentlyContinue).TcpTestSucceeded } }
```

* `True` on every line: the machine reaches the platform. If the agent is still not connected, check SSL inspection and the site token ([Installed but not connected](#installed-but-not-connected)).
* `False` on at least one line: a firewall or proxy blocks that domain. Take a screenshot and go to [Installed but not connected](#installed-but-not-connected).
  {% endtab %}
  {% endtabs %}

## Run the .msi with administrator rights

Use this if Windows did not show the `Do you want to allow this app to make changes to your device?` window and the installation failed: the installer may not have had administrator rights.

1. Press the Windows key, type `cmd`, right-click `Command Prompt`, select `Run as administrator` and click `Yes`.
2. Type `msiexec /i` followed by a space. Do not press Enter yet.
3. Drag the `.msi` file from the Desktop and drop it into the black window. Its full name appears after `msiexec /i`.
4. Press Enter. The installation window opens, this time with administrator rights.
5. Accept the license agreement, paste your site token in the `Site Token` field, click `Install`, then `Finish`.

Still failing? Go to [Basic checks](#basic-checks).

## Basic checks

Go through these checks in order. After each fix, install again.

* **Restart the machine.** A pending restart is the most common cause of failure. Click the Windows Start button, then the power icon, then `Restart`. If you see `Update and restart`, choose it and wait for the updates to finish.
* **Check that the previous antivirus is uninstalled.** In `Add or remove programs`, look for a security product (Sophos, ESET, Bitdefender, Kaspersky, Avast, AVG, McAfee, Norton, Trend Micro, Webroot, Symantec, WithSecure, CrowdStrike). If you find one, uninstall it and restart. If the uninstallation is blocked or asks for a password, contact whoever manages that antivirus. Microsoft Defender is built into Windows and does not appear in this list: leave it as it is.
* **Check that your account is an administrator.** Press the Windows key, type `Control Panel`, press Enter, then `User Accounts` > `User Accounts`. If the line under your name contains `Administrator`, your account is an administrator. Otherwise, ask someone with an administrator account to log in or to type their password during the installation.

Still failing with the `.msi`? Go to [Install from the command prompt](#install-from-the-command-prompt).

## Install from the command prompt

For Intel / AMD machines, when the sections above did not help. On ARM, this is the method of the install step: if it failed, try the PowerShell commands below with the arm64 file name.

1. In `C:\`, create a folder `Temp` and, inside it, a folder `S1`. Move the installer into `C:\Temp\S1`.
2. Press the Windows key, type `cmd`, right-click `Command Prompt`, select `Run as administrator` and click `Yes`.
3. Type the commands below, pressing Enter after each line. Replace `<version>` and `<site_token>`, and remove the `<` and `>` characters.

With the `.exe` (download `SentinelOneInstaller_windows_64bit_v<version>.exe` from the `Download the executable` menu first):

```
cd C:\Temp\S1
SentinelOneInstaller_windows_64bit_v<version>.exe -t "<site_token>" -q
```

Wait about 5 minutes, then type `echo %errorlevel%` and press Enter. `0` means the installation succeeded. Any other number: see [Exit code other than 0](#exit-code-other-than-0).

If it failed with the `.exe`, try the `.msi` in the same window. This command also writes a detailed log:

```
start /wait msiexec /i "SentinelInstaller_windows_64bit_v<version>.msi" SITE_TOKEN="<site_token>" /qn /norestart /L*v "C:\Windows\Temp\S1_install.log"
echo %errorlevel%
```

<details>

<summary>Last attempt: PowerShell</summary>

PowerShell does not always work with the installer, so use it only as a final attempt. Press the Windows key, type `PowerShell`, right-click `Windows PowerShell`, select `Run as administrator` and click `Yes`. Replace `<version>` and `<site_token>`, and keep the single quotes `'` and double quotes `"` exactly as they are.

With the `.exe` (on ARM, use the arm64 file name):

```powershell
cd C:\Temp\S1
Start-Process ".\SentinelOneInstaller_windows_64bit_v<version>.exe" -ArgumentList '-t "<site_token>" -q' -Wait -PassThru | Select-Object ExitCode
```

If it fails, with the `.msi`:

```powershell
Start-Process msiexec.exe -ArgumentList '/i "C:\Temp\S1\SentinelInstaller_windows_64bit_v<version>.msi" SITE_TOKEN="<site_token>" /qn /norestart /L*v "C:\Windows\Temp\S1_install.log"' -Wait -PassThru | Select-Object ExitCode
```

Wait until a number appears under `ExitCode`. `0` means the installation succeeded. Any other number: see [Exit code other than 0](#exit-code-other-than-0).

</details>

## Installed but not connected

The agent is on the machine but does not reach the Stoïk platform. The cause is almost always the network or the site token.

* **Site token**: a site token that was cut or pasted with a space installs the agent but never connects it. If in doubt, ask Stoïk to check it.
* **Firewall or proxy**: ask whoever manages the network to allow outbound traffic on port 443 to all the domains and IP addresses in [Network requirements](https://docs.stoik.io/mdr/deploy/sentinelone/network). No inbound port is needed. If the firewall inspects HTTPS traffic (SSL inspection), it must not inspect these domains (they all end with `.sentinelone.net`).
* Restart the machine after any network change, then wait 15 minutes.

To confirm the connection problem, or to give details to Stoïk, open a command prompt as administrator, then:

```
cd "C:\Program Files\SentinelOne\
sentinelctl management status
```

Press Tab after the path to complete the folder name. Take a screenshot of the result and send it to Stoïk with the agent logs in `C:\ProgramData\Sentinel\logs`.

## Exit code other than 0

| Code          | Meaning                                              | What to do                                                                                                                                                         |
| ------------- | ---------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| 2015          | The machine does not meet SentinelOne's requirements | `C:\Windows\Temp\SC-exit-code.txt` names the failing requirement. Usual causes: a pending restart, leftovers of the previous antivirus, an outdated Windows build. |
| 2008          | The site token is missing                            | Run the command again with the site token: `-t "<site_token>"` for the `.exe`, `SITE_TOKEN="<site_token>"` for the `.msi`. Check that it was pasted in full.       |
| 100, 103, 104 | A restart is required                                | Restart the machine, then run the same command again.                                                                                                              |
| 1603          | The `.msi` installation failed                       | Open `C:\Windows\Temp\S1_install.log` and search for `Return value 3`: the lines just above give the cause.                                                        |
| 1618          | Another installation is already running              | Wait for Windows Update to finish, or restart.                                                                                                                     |
| 1663          | Wrong installer for this machine                     | Check the table in the install step and download the matching file.                                                                                                |

## Contact support

If nothing works, send the following to <protect@stoik.io>:

* A screenshot of the `About your PC` window (Windows version and processor type).
* The exact name of the file you used and the methods you tried.
* The error message or exit code, with a screenshot if possible.
* Whether another antivirus or a previous SentinelOne agent was installed, and whether a proxy or firewall filters outbound traffic.
* The logs below.

| Log                                            | Location                                                                      |
| ---------------------------------------------- | ----------------------------------------------------------------------------- |
| Installer exit code                            | `C:\Windows\Temp\SC-exit-code.txt` and `SC-after-reboot-exit-code.txt`        |
| `.msi` installation log (command prompt)       | `C:\Windows\Temp\S1_install.log`                                              |
| `.msi` installation log (double-click)         | `MSI*.log` files in `%TEMP%` (type `%TEMP%` in the File Explorer address bar) |
| Agent logs (agent installed but not connected) | `C:\ProgramData\Sentinel\logs`                                                |
| Windows events                                 | `Event Viewer` > `Windows Logs` > `Application`, source `MsiInstaller`        |

<details>

<summary>Gather the installation logs in a single file</summary>

This creates a `S1_logs.zip` file with the installation logs of the last two days.

1. Press the Windows key, type `PowerShell`, right-click `Windows PowerShell`, select `Run as administrator` and click `Yes`.
2. Copy the three lines below, then right-click once in the PowerShell window to paste them, and press Enter.

   ```powershell
   $out = "C:\Temp\S1_logs"; New-Item -ItemType Directory -Force $out | Out-Null
   Get-ChildItem "$env:TEMP","C:\Windows\Temp" -Include MSI*.log,SC-*.txt,S1_*.log -Recurse -ErrorAction SilentlyContinue | Where-Object LastWriteTime -gt (Get-Date).AddDays(-2) | Copy-Item -Destination $out
   Compress-Archive "$out\*" "C:\Temp\S1_logs.zip" -Force
   ```
3. Wait until a new line starting with `PS C:\` appears, then close the window. Red text about a file that cannot be found is not a problem.
4. Open File Explorer, type `C:\Temp` in the address bar and press Enter. Attach `S1_logs.zip` to your email.

Without PowerShell: open File Explorer, type `C:\Windows\Temp` in the address bar and press Enter. Click the `Date modified` column to show the newest files first, and copy the `SC-exit-code.txt`, `S1_install.log` and `MSI…log` files of today to the Desktop. Do the same with `%TEMP%` in the address bar. Attach these files to your email.

</details>


# macOS

Install the SentinelOne agent on a Mac manually, from the command line or through an MDM. To deploy the EDR on a large scale, start with a manual test installation on a few endpoints.

{% stepper %}
{% step %}

## Download the installer

The executable is only available from Stoïk Protect: go to `MDR` > `Settings`, click `macOS`, then click `Download the executable`. Come back here once it's downloaded to continue the installation.
{% endstep %}

{% step %}

## Find your customer ID

Go to `MDR` > `Settings` and copy your customer ID.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FemeExV0LCLaQAfG1qQgL%2Fimage.png?alt=media&amp;token=99f2d142-fd0d-4753-9c72-f1002e271f6d" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

## Install SentinelOne

{% tabs %}
{% tab title="Manual" %}

#### Deploy the installer

1. Launch the executable.
2. Enter your customer ID in the `Site token` field.
3. Click `Continue`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FJCsBp5cQmJctE8EVUNVZ%2Fuuid-4ee31d58-fe70-7b9c-2e98-4c39e38bf7f7-png.webp?alt=media&amp;token=a669f3b1-79a8-4cd9-901c-33cfea9f6cfe" alt=""><figcaption></figcaption></figure>

#### Authorize Full Disk Access

Required if SentinelOne prompts you to authorize Full Disk Access after installation.

1. Go to `System Settings` > `Privacy & Security` > `Full Disk Access`.
2. Click the `+` button.
3. Press `Command+Shift+G` to open the `Go to Folder` menu.
4. Enter the path:

```
/Library/Sentinel/sentinel-agent.bundle/Contents/MacOS/
```

5. Double-click the destination folder.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FZPSxqtZAL3ovsGwAU4xu%2Fuuid-2e6bf2d7-20a9-a05e-f650-20140a54ed5d-png.webp?alt=media&amp;token=8a39116a-17e5-4107-a6f9-98355999a8aa" alt=""><figcaption></figcaption></figure>

6. Select the following applications and click `Open`:

   * `sentineld.app`
   * `sentineld_helper.app`
   * `sentineld_shell.app`

   If `sentineld_shell.app` does not appear in `Full Disk Access`, add it manually: click `+`, go to the same folder and select it.
7. Close the window once they appear as activated in `Full Disk Access`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F9tGRTvNMhitgy2dR7Eou%2Fuuid-7d0d2299-98ba-36a6-c52b-a989a1b7d26f-png.webp?alt=media&amp;token=2774e998-796d-4011-ad6e-bc7b9c7a7567" alt=""><figcaption></figcaption></figure>
{% endtab %}

{% tab title="Command-line" %}

1. Save your customer ID to a file at: `com.sentinelone.registration-token`. Ensure it belongs to the root user:

```bash
sudo chown root com.sentinelone.registration-token
```

2. Run the installer, replacing the path with your downloaded executable:

```bash
sudo /usr/sbin/installer -pkg /tmp/SentinelXXXX.pkg -target /
```

{% endtab %}

{% tab title="MDM" %}
The following guides detail the installation steps and prerequisites for deploying SentinelOne via MDM:

* [Jamf deployment guide](https://hcsonline.com/images/PDFs/Jamf_Install_SentinalOne.pdf)
* [Guardz / Jamf deployment guide](https://support.guardz.com/en/articles/10324446-installing-sentinelone-macos-agents-with-jamf)
  {% endtab %}
  {% endtabs %}
  {% endstep %}
  {% endstepper %}


# Linux

Install the SentinelOne agent on a Linux endpoint with the RPM or DEB package. To deploy the EDR on a large scale, start with a manual test installation on a few endpoints.

{% stepper %}
{% step %}

## Download the installer

The package is only available from Stoïk Protect: go to `MDR` > `Settings`, click `Linux`, then click `Download the executable`. Come back here once it's downloaded to continue the installation.
{% endstep %}

{% step %}

## Find your customer ID

Go to `MDR` > `Settings` and copy your customer ID.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FemeExV0LCLaQAfG1qQgL%2Fimage.png?alt=media&amp;token=99f2d142-fd0d-4753-9c72-f1002e271f6d" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

## Install SentinelOne

1. Install the package:

{% tabs %}
{% tab title="RPM" %}

```bash
sudo rpm -i <package_name>
```

{% endtab %}

{% tab title="DEB" %}

```bash
sudo dpkg -i <package_name>
```

{% endtab %}
{% endtabs %}

2. Set your customer ID:

```bash
sudo /opt/sentinelone/bin/sentinelctl management token set <customer_ID>
```

3. Start the agent:

```bash
sudo /opt/sentinelone/bin/sentinelctl control start
```

4. After a few minutes, verify the agent is running:

```bash
sudo /opt/sentinelone/bin/sentinelctl control status
```

{% endstep %}
{% endstepper %}


# Add-on modules

Set up the optional modules available with SentinelOne. To buy an additional option, contact us at <protect@stoik.io>.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-id-badge">:id-badge:</i></h4></td><td><h4>Identity Security</h4></td><td>Protect Active Directory and Entra ID with ISPM, ISIDP and decoys.</td><td><a href="https://docs.stoik.io/mdr/deploy/sentinelone/add-on/identity-security">Identity Security</a></td></tr></tbody></table>


# Identity Security

Set up Identity Security, SentinelOne's unified identity protection SKU, on your Active Directory. It covers three engines: ISPM (Active Directory posture assessment), ISIDP (real-time protection of Domain Controllers), and IDR (decoys on endpoints). Microsoft Entra ID integration is optional.

All paths below refer to the Singularity Operations Center sidebar (`Policies and settings` is the button at the bottom of the sidebar).

## Prerequisites

* **Licenses**: Endpoint (Core, Control or Complete) + **Identity Security**, on the Account *and* on every Site.
* **Windows agents**: **≥ 25.2.3.437** on DCs, servers and the AD Connector server. Workstations (IDR): ≥ 24.1.5.277.
* **Legacy Identity Agent**: DCs already on Windows agent 25.2: nothing to do. Only if the legacy Identity Agent is still present: uninstall it, then reboot (never alongside the Windows agent).
* **AD Connector server**: a domain member server, see [Choose the AD Connector server](#choose-the-ad-connector-server).
* **AD Connector → DC flows**: 53 · 88 · 389/636 · 445 · 5986 (WinRM HTTPS) · RPC Event Log.
* **Outbound flows**: HTTPS 443 / WebSocket to the S1 console for your region.
* **AD service account**: a read-only standard user, see [Create the AD service account](#create-the-ad-service-account).
* **WinRM HTTPS on every DC**:

```
winrm quickconfig -transport:https     :: WinRM HTTPS (certificate issued to the DC FQDN)
winrm configsddl default               :: Read + Execute for the service account
:: Firewall: allow TCP 5986 + "Remote Event Log Management (RPC / RPC-EPMAP)"
```

## Enable the Identity engine

`Policies and settings` > `Endpoint/Cloud Workload Security` > `Policy` > `Detection Engines`

1. Select the Account scope (or Site for an MSSP) in the scope selector at the top, then enable **Identity Detection & Response**.
2. Keep the default Identity Settings, with the Update interval set to at least 60 min.
3. In `Policies and settings` > `Accounts`, set **Identity Security Posture Mode** to **Full**.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-ff47e3c5a07c99859df109907264e4d57060c7b6%2Fs1-identity-security-detection-engines-enabled.png?alt=media" alt=""><figcaption><p>Detection Engines > Identity Detection &#x26; Response.</p></figcaption></figure>

## Connect Active Directory

`Policies and settings` > `Identity Security` > `Active Directory` > `Connect AD`

Examples in this section use a fictitious forest `acme.local`, with the child domain `corp.acme.local` (NetBIOS `CORP`).

### Create the AD service account

ISPM, ISIDP and IDR only need **read access**. A Domain Admin account is unnecessary and not recommended. ISPM remediation scripts are run separately, manually, by a Domain Admin.

<table data-search="false"><thead><tr><th>Requirement</th><th>Why</th></tr></thead><tbody><tr><td>Dedicated <strong>standard</strong> user (e.g. <code>svc-s1-identity</code>)</td><td>Read-only is enough, and ensures traceability</td></tr><tr><td>Long, random password that <strong>never expires</strong></td><td>If the password changes in AD, the connection drops (status <em>Disabled</em>)</td></tr><tr><td>Created in the forest <strong>root domain</strong></td><td>Required to assess the whole forest with Monitor subdomains</td></tr><tr><td>Kerberos <strong>AES 128 and AES 256</strong> checked</td><td>Required for assessment</td></tr><tr><td>Member of <strong>Remote Management Users</strong></td><td>WinRM access to DCs (threats and some exposures)</td></tr><tr><td>Member of <strong>Event Log Readers</strong></td><td>Reads DC Security logs (source IP and user in detections)</td></tr><tr><td>Read on adminSDHolder, KRBTGT, Administrator, Domain / Schema / Enterprise Admins, MicrosoftDNS container</td><td>Otherwise the "Insufficient permissions for AD exposure assessment" exposure is raised</td></tr><tr><td>Read on <code>CN=Deleted Objects</code></td><td>Detects mass account deletion</td></tr></tbody></table>

<details>

<summary>PowerShell example</summary>

To run on a DC or with RSAT, adjusting names as needed:

```powershell
# 1. Create the account
New-ADUser -Name "svc-s1-identity" -SamAccountName "svc-s1-identity" `
  -UserPrincipalName "svc-s1-identity@acme.local" -Path "OU=Service Accounts,DC=acme,DC=local" `
  -AccountPassword (Read-Host -AsSecureString "Password") -Enabled $true `
  -PasswordNeverExpires $true -CannotChangePassword $true `
  -KerberosEncryptionType AES128,AES256

# 2. Built-in groups
Add-ADGroupMember -Identity "Remote Management Users" -Members "svc-s1-identity"
Add-ADGroupMember -Identity "Event Log Readers"       -Members "svc-s1-identity"
```

```
:: 3. Read rights (List Contents + Read Property), inherited
dsacls "DC=acme,DC=local" /I:T /g ACME\svc-s1-identity:LCRP
dsacls "CN=Deleted Objects,DC=acme,DC=local" /takeownership
dsacls "CN=Deleted Objects,DC=acme,DC=local" /I:T /g ACME\svc-s1-identity:LCRP

:: 4. On EVERY DC: WinRM access and Security log read for Network Service
winrm configsddl default                       :: add the account, check Read + Execute
wevtutil gl security                           :: output must contain (A;;0x1;;;S-1-5-20)
wevtutil sl security /ca:"<current SDDL>(A;;0x1;;;S-1-5-20)"
```

</details>

With *Monitor subdomains*, repeat the `dsacls` commands in every domain where default permissions have been restricted.

### Choose the AD Connector server

* Domain **member** server, not a DC. Example: `srv-s1adc01.corp.acme.local`.
* Always on and not network-quarantined, 4 vCPU / 16 GB RAM, Windows agent **≥ 25.2.3.437**.
* It must **resolve via DNS** and reach every declared domain on ports 53, 88, 389/636, 445 and 5985/5986.
* A single forest, even with multiple domains: **one connector** is enough. Isolated forests without trust: one connector per forest.

Quick checks to run from this server:

```powershell
Resolve-DnsName corp.acme.local
Test-NetConnection dc01.corp.acme.local -Port 636     # LDAPS
Test-NetConnection dc01.corp.acme.local -Port 5986    # WinRM HTTPS
```

### Complete the Connect AD wizard

1. **AD Connector**: type at least 3 characters of the hostname (e.g. `srv`), pick the endpoint with a green status, then *Next*. A banner confirms the connection after a few minutes.
2. **Domain**: fill in the fields from the table below, then *Next*. The console runs an LDAP test, and the domain is saved only if the test succeeds.
3. **Decoys**: configure the decoys, see [Decoys](#decoys), then *Finish*.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-2948384debcb6e8660f9c15c95d090e84b53cdb7%2Fs1-identity-security-connect-ad-domain-form.png?alt=media" alt=""><figcaption><p>Domain step of the wizard. In this documentation example, LDAP and WinRM HTTP are selected; in production, prefer LDAPS and WinRM HTTPS.</p></figcaption></figure>

<table data-search="false"><thead><tr><th>Field</th><th>What to enter</th><th>Example</th></tr></thead><tbody><tr><td><strong>Site</strong></td><td>At Account scope: the sites that inherit this domain (useful for IDR and ISIDP). "All" does not cover sites created later.</td><td>Paris, Lyon</td></tr><tr><td><strong>Domain name</strong></td><td>The domain's <strong>DNS FQDN</strong>, never the NetBIOS name. To cover the whole forest, enter the root domain.</td><td><code>acme.local</code> or <code>corp.acme.local</code>, not <code>CORP</code></td></tr><tr><td><strong>Monitor subdomains</strong></td><td>Assesses every domain in the forest (same tree and other trees). Without this box, subdomains are discovered but <strong>not assessed</strong>.</td><td>Checked</td></tr><tr><td><strong>Domain Controller FQDN</strong></td><td>The FQDN of <strong>one</strong> DC in the domain, resolvable by the AD Connector. An RODC is accepted. No IP address. Other DCs are discovered automatically (PDC first).</td><td><code>dc01.corp.acme.local</code>, not <code>10.0.0.10</code></td></tr><tr><td><strong>Domain username</strong></td><td>The service account, in UPN, NetBIOS or plain-name format</td><td><code>svc-s1-identity@acme.local</code>, <code>ACME\svc-s1-identity</code>, or <code>svc-s1-identity</code>; not <code>ACME.LOCAL\svc-s1-identity</code></td></tr><tr><td><strong>Domain password</strong></td><td>Enter carefully: a wrong password <strong>locks the account</strong> after repeated attempts</td><td></td></tr><tr><td><strong>Refresh schedule</strong> (day, time, time zone)</td><td>AD data refresh, to schedule during off-peak hours. It does not trigger the ISPM assessment, which stays fixed at Saturday 04:00 UTC.</td><td>Sunday · 02:00 · Europe/Paris</td></tr></tbody></table>

### Protocols and options

<table data-search="false"><thead><tr><th width="140">Option</th><th width="160">Choice</th><th>Prerequisites and consequences</th></tr></thead><tbody><tr><td rowspan="2"><strong>LDAP method</strong></td><td><strong>LDAPS (636)</strong>, recommended</td><td>Encrypts traffic between the connector and the DCs. Requires a <strong>valid server certificate on the DCs</strong> (AD CS or a "Kerberos / Domain Controller Authentication" certificate) and port 636 open. Check: <code>ldp.exe</code>, SSL connection to <code>dc01.corp.acme.local:636</code>.</td></tr><tr><td><strong>LDAP (389)</strong></td><td>Unencrypted channel, reserved for DCs without a certificate or for troubleshooting. The "LDAPS connection failed" error prompts you to retry over LDAP. Fixed ports, not configurable.</td></tr><tr><td rowspan="2"><strong>WinRM method</strong></td><td><strong>HTTPS (5986)</strong>, recommended</td><td>Requires an HTTPS listener on <strong>every DC</strong>, with a certificate whose Subject is the DC FQDN (the LDAPS certificate works), and port 5986 open: <code>winrm quickconfig -transport:https</code>.</td></tr><tr><td><strong>HTTP (5985)</strong></td><td>If DCs have no certificate. Authentication remains Kerberos, but <code>AllowUnencrypted=true</code> is required on both client and service. <strong>Without WinRM on all DCs</strong>, Threat Detection fails and some exposures become <em>Skipped</em>.</td></tr><tr><td><strong>LDAP referral</strong></td><td>Checked if the forest has subdomains</td><td>Follows referrals to subdomains: ISIDP and IDR policies then also apply to child domains, which become visible in Protocol Configuration. The account must be able to query these subdomains. If enabled afterwards, run <em>Refresh AD Data</em>.</td></tr><tr><td><strong>AD Sync</strong></td><td>Checked</td><td>Syncs user and group attributes. Required for Threat Detection, CCP, compromised account scanning and asset inventory. The first sync can take a while.</td></tr><tr><td><strong>Threat Detection</strong></td><td>Checked (requires AD Sync)</td><td>Real-time ISPM threats: password spray, mass lockouts and resets, etc. Relies on a persistent LDAP connection to the PDC and on WinRM. Disabling it deletes the synced objects.</td></tr><tr><td><strong>Access over trust</strong></td><td>Only for another trusted forest</td><td>To assess a domain in another forest using the main forest's account. Two extra fields: <em>Trust domain name</em> and <em>Trust domain controller</em> (FQDN, no IP). <strong>Cannot be edited</strong> afterwards: delete and recreate.</td></tr></tbody></table>

<details>

<summary>Access over trust example</summary>

The `filiale.local` forest trusts `acme.local`. Enter Domain name = `filiale.local`, Username = `svc-s1-identity@acme.local`, Access over trust = checked, Trust domain name = `acme.local`, Trust domain controller = `dc01.acme.local`. The connector must resolve and reach the `filiale.local` DCs.

</details>

### Decoys

IDR uses these decoys.

<table data-search="false"><thead><tr><th>Field</th><th>Advice</th><th>Example</th></tr></thead><tbody><tr><td>Country for name generator</td><td>Country consistent with your users</td><td>France</td></tr><tr><td>Service (SPN) prefix</td><td>The prefix of your real service accounts, if any</td><td><code>svc</code></td></tr><tr><td>Name format</td><td>Same format as your real usernames. Variables: <code>%FN%</code> (first name), <code>%LN%</code> (last name), <code>%FNI%</code> and <code>%LNI%</code> (initials), <code>%#n%</code> (n random digits).</td><td><code>%FNI%.%LN%</code> → <code>j.dupont</code></td></tr><tr><td>Hostname format</td><td>Modeled on your workstation and DC naming</td><td>Custom: <code>PC-</code> + digits</td></tr><tr><td>IP Address format</td><td>Range or CIDR of <strong>more than 250 addresses</strong>, <strong>never used in production</strong> (otherwise false positives)</td><td><code>10.250.0.0/24</code></td></tr></tbody></table>

{% hint style="warning" %}
Decoys can no longer be edited while an active Identity protection policy covers the domain.
{% endhint %}

### After validation

* The domain status changes to **Enabled**. It reverts to *Disabled* if the connector goes down or the account password changes.
* The console runs AD learning, sync, then the ISPM assessment (about 30 min). Multiple domains added at once are processed one after another.
* Default policies are created, **inactive**.
* Each domain can be declared only **once per organization**. After deleting one, wait 120 s before recreating it. After adding a DC, click **Refresh AD Data**, otherwise DCSync false positives appear.

If the domain cannot be saved, see [Troubleshooting](#troubleshooting).

## Review the AD posture (ISPM)

`Policies and settings` > `Identity Security` > `Identity posture settings`

Nothing else to do: the assessment starts as soon as the domain is saved (about 30 min), then runs every Saturday at 04:00 UTC.

1. For an immediate scan: **Scan Now**.
2. To review results: `Misconfigurations` > `Identity`. Address Critical and High exposures first (click the misconfiguration, then `Actions` > `Mitigate` > `Download Script`).

## Protect Domain Controllers (ISIDP)

`Policies and settings` > `Identity Security` > `Identity Policies` (`Identity provider policy` / `Conditional Access policy` tabs)

Two default policies were created when the domain was saved, **inactive and read-only**. Duplicate them:

1. **Conditional Access policy**: duplicate `Default_policy_account_ca`, add an **Attack based** rule covering all attacks with the **Audit** action, then activate it.
2. **Identity provider policy**: duplicate `Default_policy_account_idp` and check the following:
   * All domains are selected and **Apply to Windows servers** is checked.
   * PAC validation stays enabled, unless a PAM or JIT solution is in place. Learning mode: enabled, 4 days.
   * Kerberos, LDAP, SAMR and LSARPC protocols checked. Add-on: *Pin version* at N-1.
   * Recommended: *Compromised credential configuration* enabled (breach databases, no mitigation at first).
   * The CA policy from step 1 is selected.
3. `Actions` > `Activate Policy`. To push immediately: sidebar `Sentinels` > `Endpoints`, select the DCs, then `Actions` > `Agent Actions` > `Push Identity Policy Changes`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-9a8a118214ab166837d7191294e7b3f153795fc2%2Fs1-identity-security-identity-provider-policy.png?alt=media" alt=""><figcaption><p>Identity provider policy: detection and protocols.</p></figcaption></figure>

{% hint style="info" %}
After a few days without false positives, switch attack rules from *Audit* to **Block**.
{% endhint %}

## Deploy decoys on endpoints (IDR)

`Policies and settings` > `Identity Security` > `Identity Policies` > `Identity protection policy`

1. Click **New Identity protection policy**, with the AD PowerShell response set to **Alert**.
2. Add 4 rules: Privileged Administrators, Service Accounts, Domain Controllers, Delegated Admins. Select the real objects, then check **Add Decoys** and **Hide Real Objects**.
3. *Create*, then `Actions` > `Activate Policy`.

Test from a domain workstation: `net group "Domain Admins" /domain` should return a decoy account. With a 25.2.3+ agent, the detection shows up as an indicator in `Event Search` (query: `indicator.name = 'IdentityDeceptivePing'`). Other alerts are in `Alerts`.

## Connect Microsoft Entra ID (optional)

Only if the customer uses Entra ID: `Policies and settings` > `Cloud accounts` > `Microsoft Azure`. You need a tenant **Global Admin** account and the Tenant ID. Optional: the Object ID of a "risky users" group.

1. *Connect account*: type **Tenant**, check **Enable Identity Security**.
2. Connection name, Tenant ID, and **Sync Microsoft Entra ID user accounts** enabled.
3. In Azure: *Tenant Properties* > *Access management for Azure resources = Yes*, then sign out and back in.
4. Download the Terraform template and run it in Azure Cloud Shell:

```
terraform init && terraform apply -auto-approve
```

5. Paste `application_id` and `client_secret`, then **Select All** and automatic onboarding, and *Finish*.

{% hint style="warning" %}
The secret **expires after one year**: plan its renewal.
{% endhint %}

Results appear in `Misconfigurations` > `Identity`.

## Verify that an alert is raised

From a domain-joined workstation, with a domain account, preferably in a pilot.

1. To see results right away: in the Identity provider policy, **temporarily disable Learning mode** (during learning, only High alerts are raised). Then push the policy to the DCs.
2. Run the commands:

```
net group "Domain Admins" /domain      :: ISIDP (SAMR) + IDR (decoy in the response)
dsquery * -filter "(admincount=1)"      :: ISIDP (LDAP), requires RSAT AD Tools
```

3. Check the `Alerts` sidebar (period *Last 1 hour*), **Detection Engine** filter:
   * `ADSecure-DC`, `-SAMR` or `-LDAP` for ISIDP (privileged account enumeration).
   * `ADSecure-EP` for IDR. With a 25.2.3+ agent, look for the indicator in `Event Search` instead.
4. Re-enable Learning mode once the test is validated.

## Final check

* The AD Connector shows an active status in `Policies and settings` > `Identity Security` > `Active Directory`.
* ISPM misconfigurations appear in the **Misconfigurations** sidebar, Identity tab.
* The Conditional Access policy and the Identity provider policy are active.
* The Identity protection policy is active.
* The `Dashboards` > `Identity` dashboard shows data, and alerts are visible in `Alerts` > `Identity`.
* AD and Entra assets are visible in `Inventory` > `Identity`.

## Troubleshooting

<details>

<summary>Save error: Authentication failed</summary>

Check the account and password (and any lockout).

</details>

<details>

<summary>Save error: LDAPS connection failed</summary>

Check the DC certificate and port 636, or switch back to LDAP.

</details>

<details>

<summary>Save error: The user account does not have the permissions…</summary>

Review the rights in [Create the AD service account](#create-the-ad-service-account).

</details>

<details>

<summary>Save error: Domain Controller down / not available / busy</summary>

Check DNS and flows from the AD Connector, then retry.

</details>


# Network requirements

Open outbound port **443** to the following domains and IP addresses if outbound traffic is blocked. They apply to every operating system, not only Windows. No inbound ports are required: all communication is outbound to SentinelOne.

## Domains

* euce1-100.sentinelone.net
* euce1-102.sentinelone.net
* euce1-103.sentinelone.net
* euce1-104.sentinelone.net
* euce1-105.sentinelone.net
* euce1-106.sentinelone.net
* euce1-108.sentinelone.net
* euce1-109.sentinelone.net
* euce1-120-mssp.sentinelone.net
* euce1-ir.sentinelone.net
* euwe3-801.sentinelone.net

## IP addresses

* 52.28.96.109
* 18.195.251.162
* 3.124.247.29
* 18.157.202.57
* 52.29.133.222
* 18.185.1.205
* 18.197.98.151
* 3.126.198.254
* 3.126.86.58
* 3.125.144.61
* 3.126.137.49
* 3.123.155.200
* 18.158.53.176
* 3.65.44.11
* 3.66.206.10
* 3.124.180.157
* 3.68.134.166
* 3.126.19.109
* 3.73.61.193
* 3.65.215.106
* 3.125.103.71
* 52.28.140.160
* 3.79.163.229
* 3.79.197.235
* 52.28.194.120
* 3.79.144.24
* 3.64.20.194
* 18.158.224.145
* 52.29.182.176
* 3.73.69.228
* 52.28.5.228
* 34.36.224.38


# Microsoft Defender

Connect your Microsoft Defender XDR / Defender for Endpoint tenant to Stoïk so the Stoïk CERT team can monitor and respond on it, as part of the MDR (Managed Detection and Response) offering. Unlike CrowdStrike or SentinelOne, Stoïk does not sell Microsoft Defender licences: the licences remain yours, and you stay in control of your Microsoft tenant.

{% hint style="warning" %}
The management of MS Defender EDR is currently in Beta. Let us know if you have any feedback or encounter any issues.
{% endhint %}

Your tenant is connected to Stoïk in two ways, and both are required:

* **The Stoïk application** ([Part A](#part-a-authorise-the-stoik-application)): sends your alerts and the list of protected devices to Stoïk Protect.
* **The Stoïk MDR guest account** ([Part B](#part-b-give-the-stoik-cert-team-access-to-your-defender-portal)): lets the Stoïk CERT team investigate and respond in your Defender portal (device timeline, isolation, live response).

The setup is complete only when Stoïk can see all your protected devices in the Defender portal: see [Check the setup and notify Stoïk](#check-the-setup-and-notify-stoik).

## Prerequisites

* The licence scope is agreed with your Stoïk sales representative, as part of the Stoïk MDR purchase. Stoïk does not purchase Microsoft licences on your behalf.
* Workstations have Defender for Endpoint Plan 2 (included in Microsoft 365 E5 / A5, and in Defender for Business for SMBs).
* A **Global Administrator** account of the tenant that holds your Defender devices. For [B.3](#b.3-create-the-stoik-role-in-defender), a **Security Administrator** account also works.

## Deploy Defender on your devices

Assigning a licence is not the same as deploying the Defender for Endpoint agent. Every device must be onboarded.

* **Workstations:** onboard them through Intune, Group Policy, or the local script. See Microsoft's [Onboarding devices](https://learn.microsoft.com/en-us/defender-endpoint/onboarding) documentation for the available methods.
* **Servers**, one of two options:
  * A Defender for Endpoint server licence: subscribe it first as part of your licence scope, see [Prerequisites](#prerequisites), then onboard the servers with Microsoft's onboarding package.
  * **Microsoft Defender for Servers** (Defender for Cloud): in `Defender for Cloud` > `Environment settings` > your subscription > `Defender plans` > `Servers` > `Settings`, set **Endpoint protection** to **On**. Connect on-premises servers through **Azure Arc** first. The Azure subscription must belong to the **same Microsoft tenant** that you connect to Stoïk, otherwise the servers appear in another tenant's console.
* **Windows Server 2012 R2 / 2016:** use the modern unified agent. Devices still onboarded through the Microsoft Monitoring Agent (MMA) must be migrated: MMA stops working on 26 February 2027. See Microsoft's [migration guide](https://learn.microsoft.com/en-us/defender-endpoint/update-agent-mma-windows).

{% hint style="warning" %}
A device appearing under `MDR` > `Monitored Assets` in Stoïk Protect does not mean Defender for Endpoint is installed and active on it. Assets shown there come from Entra ID, not from Defender itself, so a device can appear before the agent is actually deployed, and servers that are not Entra-joined may be missing. The reference list is `Assets` > `Devices` in the Defender portal.
{% endhint %}

## Part A: Authorise the Stoïk application

Authorise the Stoïk application in a single click from Stoïk Protect.

{% hint style="warning" %}
Do Part A and Part B in the **same Microsoft tenant**: the one where your devices appear under `Assets` > `Devices` in the Defender portal.
{% endhint %}

1. Log in to [Stoïk Protect](https://app.stoik.io) with your Stoïk account.
2. Go to `MDR` > `Settings`.
3. Click **Connect your Microsoft account**.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FJ27yVAk93QDQsSIRVQsx%2FScreenshot%202026-04-21%20at%2010.11.45.png?alt=media&amp;token=ffa7861f-44e1-498c-a111-eaa8562a86ae" alt=""><figcaption></figcaption></figure>
4. You are redirected to Microsoft. Review the permissions requested by the Stoïk MDR application and click **Accept**.
5. You are sent back to Stoïk Protect. The **Settings** page shows the connection as **Connected**, along with the detected **Microsoft Tenant ID**. Write it down: it must be the same tenant ID as the one you copy in step [B.1](#b.1-invite-the-stoik-mdr-user). If it differs, you connected another tenant: contact <protect@stoik.io> before going further.

## Part B: Give the Stoïk CERT team access to your Defender portal

Do all four steps, B.1 to B.4, whatever your Defender configuration. Microsoft tenants use different permission models: default, legacy Defender for Endpoint roles, or Defender XDR Unified RBAC (the default for tenants new to Defender for Endpoint since 2025). These steps work in all of them, and a step that is not needed in your tenant does no harm.

### B.1: Invite the Stoïk MDR user

{% stepper %}
{% step %}
Go to [portal.azure.com](https://portal.azure.com) (or [entra.microsoft.com](https://entra.microsoft.com)) with a **Global Administrator** account.
{% endstep %}

{% step %}
Go to `Users` > `All users`, then click **Invite external user**.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FFJBh3TOnPyKo4oQgp1lp%2Fazure-portal-users-invite-external-user.png?alt=media&amp;token=082c050f-ea20-4b37-a11c-70e49f8e0043" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Fill in:

* **Email address:** `mdr_api@stoikcert.onmicrosoft.com`
* **Display name:** `Stoik_Mdr_Api`

Copy your **tenant ID** (displayed at step 3 of the invite form). It must match the tenant ID shown in Stoïk Protect in Part A.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FmMCAn1mu4QB71K9WwsYG%2Fazure-portal-invite-tenant.png?alt=media&amp;token=d8dab891-9674-4e9b-a13e-8097611752fc" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Go to the **Assignments** tab, click **Add role**, search for and select **Security Operator**, then click **Select**.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F2oHLYYIQa0PUzkFIbIDM%2Fazure-portal-invite-external-user-assignments.png?alt=media&amp;token=487ffb49-f790-41f0-bd9f-61994b7a8713" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
If the role does not appear in the assignments list afterward, this is a known Microsoft display bug. Refresh the page, the role is usually applied correctly.
{% endhint %}
{% endstep %}

{% step %}
Click **Review + invite**, then **Invite** to finalize.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fltyza0X2WhCcXbIErXpw%2Fazure-portal-invite-external-user-review.png?alt=media&amp;token=1c0cb9df-8683-4ccf-b817-92fc7a4bf5e9" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}
{% endstepper %}

{% hint style="info" %}
In your tenant, the guest's user name becomes `mdr_api_stoikcert.onmicrosoft.com#EXT#@<your-tenant>.onmicrosoft.com`. In the next steps, search for it by its display name **Stoik\_Mdr\_Api**, not by its email address.
{% endhint %}

<details>

<summary>A Conditional Access policy blocks the invited user</summary>

If your tenant has Conditional Access policies in place, they may block the Stoïk MDR guest account. Exclude the `Stoik-MDR` group, created in [B.2](#b.2-create-the-stoik-mdr-group), from the blocking policy:

1. Go to `Protection` > `Conditional Access` > `Policies`.
2. Click the policy that is blocking access.
3. Under the `Users` section, click the `Exclude` tab, tick `Users and groups`, then search for and select the `Stoik-MDR` group.
4. Click **Save**.

</details>

<details>

<summary>Your tenant enforces MFA</summary>

If your tenant enforces MFA, the Stoïk MDR user will be prompted for it, which the Stoïk CERT team cannot complete. Instead of exempting the account from MFA, set up a cross-tenant trust so the MFA already performed on Stoïk's tenant is accepted by yours:

1. Sign in to the [Microsoft Entra admin center](https://entra.microsoft.com).
2. Go to `Identity` > `External Identities` > `Cross-tenant access settings`.
3. Under the `Organizational settings` tab, click **Add organization**.
4. Type Stoïk's tenant ID `8100d7bd-172a-492e-bb96-e3f923731bd9` (or `stoikcert.onmicrosoft.com`), select it, then click **Add**.
5. Once the organization appears in the list, click the link under the `Inbound access` column (typically `Inherited from default`).
6. Go to the `Trust settings` tab.
7. Select **Customize settings**.
8. Tick `Trust multi-factor authentication from Microsoft Entra tenants`.
9. Click **Save**.

</details>

### B.2: Create the Stoik-MDR group

Defender device groups and legacy Defender roles only accept groups, never individual users. Use this group, not the user, in all the following steps.

1. In [entra.microsoft.com](https://entra.microsoft.com), go to `Groups` > `All groups` > **New group**.

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-b3101e4ff5f5742aa7cc717174855a0dc5ce5b65%2Fentra-groups-new-group.png?alt=media" alt=""><figcaption></figcaption></figure>
2. Fill in:

   * **Group type:** `Security`
   * **Group name:** `Stoik-MDR`
   * **Membership type:** `Assigned`

   <figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-4714c19082562b69aaee2b2a50e08c0f0d364475%2Fentra-new-group-stoik-mdr.png?alt=media" alt=""><figcaption></figcaption></figure>
3. Under **Members**, click **No members selected**, add **Stoik\_Mdr\_Api**, then click **Create**.

### B.3: Create the Stoïk role in Defender

This step is mandatory, even if you do not know which Defender permission model your tenant uses.

{% stepper %}
{% step %}
Go to [security.microsoft.com](https://security.microsoft.com) with a **Global Administrator** or **Security Administrator** account.
{% endstep %}

{% step %}
Expand `System`, click `Permissions`, then under `Microsoft Defender XDR (Unified RBAC)`, select `Roles`.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FdJanRdyTLyPrWq2XkH5v%2Fmicrosoft-security-system-permissions-roles.png?alt=media&amp;token=b67733d5-09f0-4481-9ab6-b3965bbb6e5d" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Click **Create custom role** and fill **Role Name** with `stoik_mdr_role`.

![](https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FH3rzxW1t5X9JnDyhVXrL%2Fmicrosoft-security-permissions-roles-add.png?alt=media\&token=2f8b3a94-1749-4846-9f89-31be3bb8fcd8)

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F2h1FzJS2rhuhbO8HA9li%2Fmicrosoft-security-permissions-roles-add-setup-basics.png?alt=media&amp;token=7a1558bb-6030-4f53-895c-fdddf228eaee" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Grant permissions: for each of the three permission groups (**Security operations**, **Security posture**, **Authorization and settings**), select **All read and manage permissions**, click **Apply**, then **Next**.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FP6xHDJ63Yb3DPCFZk3oE%2Fmicrosoft-settings-authorization-and-settings.png?alt=media&amp;token=98c0fe8e-133b-41e0-85d3-2dbca7e2a570" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Add the assignment:

* Click **Add assignment**.
* Name: `stoik_mdr_assignment`.
* Assign to: the **`Stoik-MDR`** group.
* Select **All data sources**.
* Tick **Include future data sources automatically**.
* Click **Add**, then **Next**.

![](https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fb3lIdpsPj6FsV0ZSMlJ7%2Fmicrosoft-security-assigment-details.png?alt=media\&token=e276326b-6c53-44c6-9b74-46377f8f18b0)
{% endstep %}

{% step %}
Click **Submit** to finalize.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F8LvLf582dkw7KpjNnqLP%2Fmicrosoft-security-assigment-validating.png?alt=media&amp;token=18d03c49-3cdc-41e1-914a-e343965e5d2e" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Still in `System` > `Permissions` > `Microsoft Defender XDR (Unified RBAC)`, open **Workload settings** (or **Activate workloads**) and check that **Microsoft Defender for Endpoint** is **activated**. If it is not, the role above is ignored for devices.
{% endstep %}
{% endstepper %}

<details>

<summary>You prefer to keep the legacy Defender for Endpoint roles</summary>

Instead of activating the Defender for Endpoint workload in Unified RBAC:

1. Go to `Settings` > `Endpoints` > `Permissions` > `Roles`.
2. Click **Add item** and name it `stoik_mdr_role`.
3. Tick **all permissions**, including **Live response advanced**.
4. In **Assigned user groups**, add `Stoik-MDR`.
5. Click **Save**.

</details>

### B.4: Give access to all your device groups

Skip this step only if you have never created a device group. A device group with access restricted to other groups hides all of its devices from Stoïk, including servers.

1. Go to `Settings` > `Endpoints` > `Permissions` > `Device groups`.
2. For **each** device group, **including `Ungrouped devices (default)`**: open it, go to the **User access** tab, add `Stoik-MDR`, then click **Save**.

Changes can take a few hours to apply.

## Check the setup and notify Stoïk

1. In [security.microsoft.com](https://security.microsoft.com), go to `Assets` > `Devices` and note your number of devices, and of servers (filter `OS platform` on Windows Server and Linux).
2. In Entra, go to `Enterprise applications` > the Stoïk application > `Permissions`, and check that all permissions show **Granted for \<your organisation>**.
3. Send an email to <protect@stoik.io> with your **Microsoft Tenant ID** and the device and server counts from point 1.

The Stoïk CERT team then logs in with the guest account and validates the onboarding only when it sees `Assets` > `Devices` with the same counts. If the menu is missing or the list is empty, see [Troubleshooting](#troubleshooting).

<details>

<summary>Re-approving the Stoïk application (existing clients)</summary>

When Stoïk adds a permission to its application (for example to show your real Defender device list in Stoïk Protect, or to isolate a device), a Global Administrator of your tenant must approve it again: open the link sent by your Stoïk CSM and click **Accept**.

This only affects the Stoïk application (Part A). It does not change the guest account's access to your Defender portal (Part B).

</details>

## Troubleshooting

<details>

<summary>Stoïk sees alerts but no Assets > Devices menu, or an empty device list</summary>

Unified RBAC or legacy roles are active and the guest has no Defender role, or a device group restricts access. Complete steps [B.2](#b.2-create-the-stoik-mdr-group) to [B.4](#b.4-give-access-to-all-your-device-groups): role assigned to `Stoik-MDR`, Defender for Endpoint workload activated, group added to every device group.

</details>

<details>

<summary>Workstations are visible, servers are missing</summary>

The servers are in a restricted device group, the Defender for Servers integration is off, or the Azure subscription is in another tenant. See [B.4](#b.4-give-access-to-all-your-device-groups) for their device group, and the servers section of [Deploy Defender on your devices](#deploy-defender-on-your-devices).

</details>

<details>

<summary>The guest account cannot sign in</summary>

A Conditional Access policy or an MFA requirement blocks it. See the two dropdowns under [B.1](#b.1-invite-the-stoik-mdr-user).

</details>

<details>

<summary>The guest is not found when assigning the role</summary>

A guest's user name contains `#EXT#`, so searching by email fails. Search for **Stoik\_Mdr\_Api**, or assign the `Stoik-MDR` group.

</details>

<details>

<summary>The tenant ID in Stoïk Protect differs from the Defender console</summary>

Part A was approved in another tenant. Contact <protect@stoik.io>, then redo [Part A](#part-a-authorise-the-stoik-application) in the tenant that holds your devices.

</details>

<details>

<summary>The devices in Stoïk Protect differ from Defender</summary>

The list in Stoïk Protect still comes from Entra ID, or the application permissions were not re-approved. See `Re-approving the Stoïk application` in [Check the setup and notify Stoïk](#check-the-setup-and-notify-stoik).

</details>

<details>

<summary>The "Connect your Microsoft account" button in Stoïk Protect returns an error</summary>

Make sure you are signed into Microsoft with a **Global Administrator** account.

</details>

<details>

<summary>The connection shows as "Connected" in Stoïk Protect but no host or alert appears</summary>

It can take up to a few hours for the first data sync.

</details>

<details>

<summary>A banner mentions devices running through MMA</summary>

Older servers still use the legacy agent, which stops working on 26 February 2027. Migrate them to the unified agent: see [Deploy Defender on your devices](#deploy-defender-on-your-devices).

</details>

For any unresolved issue or to revoke Stoïk's access to your Microsoft Defender tenant, contact <protect@stoik.io>.


# Management

Monitor and manage your MDR deployment from the [Stoïk Protect platform](https://app.stoik.io): track your fleet, read results, and set who gets alerted.

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4><i class="fa-server">:server:</i></h4></td><td><h4>Hosts table</h4></td><td>Check the status and details of every endpoint monitored by the MDR service.</td><td><a href="https://docs.stoik.io/mdr/management/hosts">Hosts table</a></td></tr><tr><td><h4><i class="fa-users">:users:</i></h4></td><td><h4>Users table</h4></td><td>Review the identity risks detected on your Active Directory users.</td><td><a href="https://docs.stoik.io/mdr/management/users">Users table</a></td></tr><tr><td><h4><i class="fa-shield-halved">:shield-halved:</i></h4></td><td><h4>MDR results</h4></td><td>Read the security alerts detected on your endpoints and identities.</td><td><a href="https://docs.stoik.io/mdr/management/results">MDR results</a></td></tr><tr><td><h4><i class="fa-trash-can">:trash-can:</i></h4></td><td><h4>Uninstalling a host</h4></td><td>Remove the EDR agent from a workstation, server or mobile device.</td><td><a href="https://docs.stoik.io/mdr/management/uninstall">Uninstalling a host</a></td></tr><tr><td><h4><i class="fa-address-book">:address-book:</i></h4></td><td><h4>Priority contacts</h4></td><td>Set who our team contacts first when an alert requires your action.</td><td><a href="https://docs.stoik.io/mdr/management/priority-contacts">Priority contacts</a></td></tr></tbody></table>


# Hosts table

The `MDR` > `Monitored Assets` > `Hosts` page lists all endpoints (workstations, servers, and mobile devices) monitored by the MDR service. Each host is a device on which an EDR agent (CrowdStrike, SentinelOne, or Microsoft Defender) has been deployed.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-0f3fd4b18a821c491cd0d38912c289ea2a0b9bf9%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2015.33.07.png?alt=media" alt=""><figcaption></figcaption></figure>

## Columns

Each row in the table displays the following information:

<table><thead><tr><th width="164.3984375">Column</th><th>Description</th></tr></thead><tbody><tr><td><strong>Status</strong></td><td>Whether the host is <strong>Active</strong>, <strong>Inactive</strong>, or <strong>Uninstalling</strong>. See <a href="#statuses">Statuses</a>.</td></tr><tr><td><strong>Name</strong></td><td>The hostname of the device as reported by the EDR agent.</td></tr><tr><td><strong>Platform</strong></td><td>The operating system family (e.g. Windows, macOS, Linux).</td></tr><tr><td><strong>OS Version</strong></td><td>The specific version of the operating system running on the device.</td></tr><tr><td><strong>Last Seen</strong></td><td>The date the EDR agent last reported activity to the provider.</td></tr><tr><td><strong>First Seen</strong></td><td>The date the EDR agent was first detected on the device.</td></tr><tr><td><strong>Type</strong></td><td>The device category as reported by the EDR provider, such as Workstation, Server, Domain Controller, or Mobile.</td></tr><tr><td><strong>Email</strong></td><td>The email address associated with the device, if available.</td></tr><tr><td><strong>Username</strong></td><td>The last user who logged into the device, if available.</td></tr></tbody></table>

You can search for a specific host by name using the search bar at the top of the table. You can also filter hosts by status, platform, OS version, and type using the filter options.

## Statuses

The status of a host reflects the state of the EDR agent on the device. It combines the agent's activity status from the EDR provider with any pending actions initiated from Stoïk Protect.

<table><thead><tr><th width="175.3203125">Status</th><th>Meaning</th></tr></thead><tbody><tr><td><strong>Active</strong></td><td>The EDR agent is online and reporting normally. The device is actively monitored.</td></tr><tr><td><strong>Uninstalling</strong></td><td>An uninstallation has been triggered from Stoïk Protect. The agent is in the process of being removed. For CrowdStrike, this status is set as soon as the uninstallation process is confirmed, even if the actual removal has not yet completed on the device.</td></tr><tr><td><strong>Inactive</strong></td><td>The EDR agent has been removed or has not reported activity for 30+ days. Inactive hosts are no longer actively monitored.</td></tr></tbody></table>

## Exporting host data

You can export the full list of monitored hosts as a file. The export includes all columns except Status: name, platform, OS version, first seen, last seen, type, email, and username.

## Uninstalling an EDR agent

For hosts with an **Active** status, you can trigger an uninstallation directly from the Hosts table using the action button on the right side of the row. This feature is available for CrowdStrike and SentinelOne endpoints. See [Uninstalling a host](https://docs.stoik.io/mdr/management/uninstall).


# Users table

The `MDR` > `Monitored Assets` > `Users` page on Stoïk Protect lists all Active Directory users monitored by the MDR service. Each user is assessed for identity-related risks such as weak or aged passwords, stale accounts, and more.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-202028efad10c945dc6f767477e5baf5231753fd%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2015.26.23.png?alt=media" alt=""><figcaption></figcaption></figure>

All values displayed in this table are retrieved directly from CrowdStrike. Stoïk displays them as-is and does not calculate or modify them.

## Columns

Each row in the table displays the following information:

<table><thead><tr><th width="208.9375">Column</th><th>Description</th></tr></thead><tbody><tr><td><strong>Status</strong></td><td>Whether the user account is <strong>Active</strong> or <strong>Inactive</strong>. Inactive users are archived accounts no longer in use.</td></tr><tr><td><strong>Name</strong></td><td>The display name of the user in Active Directory.</td></tr><tr><td><strong>Risk</strong></td><td>A risk score severity level: <strong>Normal</strong>, <strong>Medium</strong>, or <strong>High</strong>, based on the combined risk factors detected for this user.<br>Users with a risk score greater than 7.5 are considered high risk, indicating a high probability of compromise.</td></tr><tr><td><strong>Attributes</strong></td><td>Special flags assigned to the user by security analysts. See <a href="#attributes">Attributes</a>.</td></tr><tr><td><strong>Risk factors</strong></td><td>The specific identity risks detected on this user. See <a href="#risk-factors">Risk factors</a>.</td></tr></tbody></table>

You can search for a specific user by name using the search bar at the top of the table.

## Attributes

Attributes are special flags used by security analysts to track specific users or set traps for attackers. They appear as icons in the Attributes column.

<table><thead><tr><th width="216.55078125">Attribute</th><th>Meaning</th></tr></thead><tbody><tr><td><strong>Marked</strong></td><td>Users singled out by security analysts for easy identification throughout the console. This flag remains active for 48 hours by default.</td></tr><tr><td><strong>Watched</strong></td><td>Users requiring special attention, such as those who have resigned, are under notice, or are suspected of being motivated to cause organizational harm.</td></tr><tr><td><strong>Honey Token</strong></td><td>Deceptive accounts used to lure attackers. Any activity or change related to a Honey Token account triggers a dedicated detection indicating malicious network activity.</td></tr></tbody></table>

## Risk factors

Risk factors are specific identity vulnerabilities detected on a user account. Each factor appears as a badge in the Risk factors column. The more risk factors a user has, the higher their overall risk score will be.

<table><thead><tr><th width="220.5078125">Risk factor</th><th>Meaning</th></tr></thead><tbody><tr><td><strong>Insufficient password rotation</strong></td><td>The user's password is set to never expire, providing attackers an unlimited timeframe to conduct brute-force attacks. It is recommended to enforce a policy requiring users to change passwords periodically.</td></tr><tr><td><strong>Aged password</strong></td><td>The user's password has not been changed for an extended period of time.</td></tr><tr><td><strong>Exposed password</strong></td><td>The user's password has been found in a known data breach or in Group Policy Preference objects, making it potentially visible to unauthorized users.</td></tr><tr><td><strong>Password brute force</strong></td><td>A brute force attack has been detected against this user's account.</td></tr><tr><td><strong>Shared user</strong></td><td>Applied to a user who logs in from multiple locations simultaneously or in a suspicious pattern.</td></tr><tr><td><strong>Golden ticket</strong></td><td>A Kerberos Golden Ticket attack has been detected, which could allow an attacker to impersonate this user indefinitely.</td></tr><tr><td><strong>Stale account</strong></td><td>An account that has been dormant or inactive for more than three months. Stale users are risky because they are rarely monitored by the owner, allowing malicious activity to go unnoticed. It is recommended to periodically review stale users and delete or disable accounts no longer in use.</td></tr><tr><td><strong>Weak password</strong></td><td>The user's password has been identified as weak or vulnerable to dictionary attacks. It is recommended to mandate the use of strong, complex passwords.</td></tr></tbody></table>


# MDR results

Review the security alerts detected on your endpoints and identities in `MDR` > `Results`, and open each one to see how it was handled.

## Results table

Each alert in the table displays:

* **Name**: the alert rule that was triggered
* **Criticality**: Low, Medium, High, or Critical
* **Target**: the affected host or user
* **Detection date**: when the threat was first detected
* **Status**: Open or Closed
* **Type**: Endpoint or Identity

## Alert details

Click any alert to open a side panel with full details:

* **Information**: detection date, affected host or user, alert type and a description of the alert when available.
* **Status**: whether the alert is Open or Closed, and its resolution (True Positive or False Positive) once closed.
* **History**: a timeline tracking the alert through its lifecycle. See [Alert history](#alert-history).
* **Linked alerts**: other alerts grouped in the same collection, if any.

## Alert history

The History section shows the key milestones of how the alert was handled:

| Step               | Meaning                                                                           |
| ------------------ | --------------------------------------------------------------------------------- |
| Alert creation     | The moment the threat was first detected by the EDR (CrowdStrike or SentinelOne). |
| Alert confirmation | When the alert was ingested and confirmed in the Stoïk system.                    |
| Alert processing   | When a SOC analyst started investigating the alert.                               |
| Alert closure      | When the alert was resolved and closed.                                           |

"Alert creation" and "Alert confirmation" are always shown. "Alert processing" and "Alert closure" appear as the alert progresses through its lifecycle. Each step displays the exact date and time, with the most recent event at the top.

## Analyst notes

Some alerts include notes written by Stoïk SOC analysts, with additional context on how the alert was investigated and resolved. The SOC team adds them only when this context helps you understand the alert resolution.

Each note displays:

* **Author**: shown as "Stoïk SOC Analyst" (individual analyst names are not disclosed)
* **Timestamp**: when the note was written
* **Content**: the analyst's observations, findings, or recommendations

## Email exchange

When emails related to the alert handling exist, they are displayed in the alert side panel, giving you the full context of how the alert was communicated and resolved. If there are none, the Email exchanges section is not displayed.


# Uninstalling a host

Remove the EDR agent from a monitored host: from Stoïk Protect on a workstation or server, or directly on a mobile device.

## Prerequisites

* **Managed services permission**: `View and edit managed services tools` must be enabled on your account. See [User access management](https://docs.stoik.io/onboarding/account/users).
* **MFA enabled for at least 2 weeks**: set it up from `User settings` > `Security`, in the menu in the lower-left corner of the app. See [Securing your account](https://docs.stoik.io/onboarding/account/mfa).

## Uninstall on a workstation or server

Go to `MDR` > `Monitored Assets` > `Hosts`. Locate the host and click the **Uninstall** icon on its row.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-e68f0edc93c96262237fea34f9f3b3f74e529f73%2FCapture%20d%E2%80%99%C3%A9cran%202026-05-07%20%C3%A0%2015.39.46.png?alt=media" alt=""><figcaption></figcaption></figure>

The next steps depend on the EDR provider running on the host.

### CrowdStrike

With CrowdStrike, the uninstall runs on the device. Clicking **Uninstall** opens a modal with a one-time **maintenance token** you'll need to remove the agent locally.

{% stepper %}
{% step %}

#### Get the maintenance token

The modal displays the token automatically. Keep this modal open during the uninstall.
{% endstep %}

{% step %}

#### Uninstall on the device

{% tabs %}
{% tab title="Windows" %}

1. Download the [**CrowdStrike Falcon uninstall tool**](https://drive.google.com/file/d/1a5RD8HvwsSl26viCYw_QWhoKRn9JfRD-/view?usp=sharing) linked in the modal.
2. Run the tool. When prompted, enter the monitored asset name and maintenance key shown in the modal.
3. Wait for the tool to confirm removal.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FtxXFyLc9VfwvO8HR9Ag0%2FScreenshot%202026-05-18%20at%2018.46.10.png?alt=media&amp;token=9658ae55-c87f-429b-b43f-7a63569b9736" alt=""><figcaption></figcaption></figure>
{% endtab %}

{% tab title="macOS" %}

1. Open Terminal and run:

```bash
sudo /Applications/Falcon.app/Contents/Resources/falconctl uninstall -t
```

2. When prompted, enter the monitored asset name and maintenance key shown in the modal.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FPiP5v7VYwNp1eyUcWVpy%2FScreenshot%202026-05-18%20at%2018.45.22.png?alt=media&amp;token=40cbe5d7-d329-4d10-8d33-8237587a46c4" alt=""><figcaption></figcaption></figure>
{% endtab %}

{% tab title="Linux" %}

1. Open a terminal.
2. Apply the maintenance token:

```bash
sudo /opt/CrowdStrike/falconctl -s --maintenance-token=<maintenance_token>
```

3. Uninstall the Falcon sensor:

   Debian / Ubuntu:

   ```bash
   sudo apt-get purge falcon-sensor
   ```

   RHEL / CentOS / Fedora:

   ```bash
   sudo yum remove falcon-sensor
   ```

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FvJzzsFTWTgYVcAEbAQP7%2FScreenshot%202026-05-18%20at%2017.55.40.png?alt=media&amp;token=4db7f351-984d-46e9-840d-b95f3d955995" alt=""><figcaption></figcaption></figure>
{% endtab %}
{% endtabs %}
{% endstep %}

{% step %}

#### Confirm in Stoïk Protect

Return to the modal, tick "I have completed the uninstallation on the device using the token above" and click **Confirm**.

The host moves to `Uninstalling`. See [After uninstallation](#after-uninstallation).

<details>

<summary>View token / Cancel uninstallation</summary>

Once an uninstall is in progress, two additional icons appear on the host's row:

* **View token**: reopens the modal in read-only mode so you can copy the maintenance token again if needed.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FlVfiiScfqZYmaiMXJHtc%2FScreenshot%202026-05-18%20at%2020.02.25.png?alt=media&amp;token=4dfde0f7-8673-4b61-91e1-dd4b9736706c" alt=""><figcaption></figcaption></figure>

* **Cancel uninstallation**: clears the `Uninstalling` status in Stoïk Protect.

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FVa3ttr1lpX3UcLYxOHtJ%2FScreenshot%202026-05-18%20at%2020.02.32.png?alt=media&amp;token=6a3ce76c-b131-470f-b88f-72def022bf4f" alt=""><figcaption></figcaption></figure>

{% hint style="warning" %}
Cancelling does not reinstall the agent on the device. If you've already run the uninstall command, the agent is already gone: cancelling simply brings the host back to `Active` in the table until CrowdStrike marks it `Inactive`.
{% endhint %}

</details>
{% endstep %}
{% endstepper %}

### SentinelOne

With SentinelOne, the uninstall is fully managed by Stoïk Protect. No command to run on the device. The flow is the same on Windows, macOS and Linux.

1. Click the **Uninstall** icon on the host's row.
2. Check the asset name in the modal and click **Confirm**. The host moves to `Uninstalling` immediately, then to `Inactive` once SentinelOne stops reporting the agent.

{% hint style="warning" %}
SentinelOne uninstalls cannot be cancelled. The order is sent to SentinelOne the moment you click **Confirm** and there is no Cancel button on the row afterwards. Double-check the host name before confirming.
{% endhint %}

## Uninstall on a mobile device

The mobile flow doesn't use Stoïk Protect. The EDR is removed directly on the device.

{% tabs %}
{% tab title="Android" %}
Remove the EDR app from the device's app management or Settings, like any other app.
{% endtab %}

{% tab title="iOS" %}
Open `Settings` > `General` > `VPN & Device Management` and delete the EDR profile, then delete the EDR app.
{% endtab %}
{% endtabs %}

## After uninstallation

The agent is removed from the device immediately, but the host record stays visible in Stoïk Protect for a while longer. How long depends on the EDR.

{% tabs %}
{% tab title="CrowdStrike" %}
The host stays in CrowdStrike for 45 days after uninstallation. During that window, its status in Stoïk Protect is `Uninstalling`. After 45 days, CrowdStrike purges the host and the status becomes `Inactive`.
{% endtab %}

{% tab title="SentinelOne" %}
The host stays in SentinelOne until the device next comes online. While it's offline, its status in Stoïk Protect is `Uninstalling`; when it reconnects, SentinelOne removes the host and the status becomes `Inactive`. If the device never reconnects (lost, stolen, decommissioned), the host stays `Uninstalling` indefinitely.
{% endtab %}
{% endtabs %}

For host statuses, see [Statuses](https://docs.stoik.io/mdr/management/hosts#statuses).


# Priority contacts

Set your priority contacts so that our team can reach you quickly when an MDR alert requires your direct action.

## When your action is required

Stoïk handles the analysis and processing of security alerts, but some situations require your direct action. For example:

* Reformatting a workstation after a confirmed infection
* Verification by the SOC: for example, if user X is using an unlisted tool Y, we will ask you to confirm its legitimacy
* Active Directory alert: if a compromised account is detected, an incident response plan must be triggered with our recommendations (network shutdown, disabling VPN accounts, etc.)

## Setting up priority contacts

Enter your priority contacts in `Settings` > `Users`. For the full setup, see [User rights and notifications](https://docs.stoik.io/onboarding/account/users#user-rights-and-notifications).


# Common questions

<details>

<summary>Does the EDR slow down my computer?</summary>

The EDR may use more resources during occasional scans, but it is generally not the cause of a prolonged slowdown. A slowdown usually comes from:

* A high workload on the computer (demanding programs, saturated memory, full disk)
* A network or server connection problem
* A background update or process
* A software malfunction

To help us find the cause, gather this information:

1. Is the slowdown **widespread**, or does it only affect one computer? If only one, note its name or identifier.
2. What is **slow**: the computer itself (applications slow to open, frozen cursor) or the connection to a server or online service?
3. Have all other **security tools** been uninstalled, in particular antivirus software and any other EDR?

Then:

1. Make sure no other antivirus or EDR is installed, and that all related services are stopped.
2. Uninstall and reinstall the EDR on the endpoint.
3. List all the running processes and services during the slowdown.

Send this information to <protect@stoik.io>. It lets us tell a local (computer), network (connection or server) or EDR problem apart.

</details>

<details>

<summary>What if Internet connectivity is lost?</summary>

As long as a device has Internet access, protection is complete and real-time: the agent communicates continuously with the cloud for threat intelligence, detection and response.

A device that is temporarily offline (travel, dead zone, etc.) remains protected:

* **Local ML models**: the agent embeds on-sensor machine learning models that detect and block threats (e.g. malware from a USB drive) without cloud connectivity.
* **No EDR capabilities**: real-time telemetry, threat hunting, and remote response actions from the console are unavailable until the device reconnects.

When the device reconnects, the telemetry recorded offline is uploaded to the console for retrospective analysis.

For private or isolated networks (air-gap), open the required network flows at the firewall: see the network requirements for [CrowdStrike](https://docs.stoik.io/mdr/deploy/crowdstrike/network) and [SentinelOne](https://docs.stoik.io/mdr/deploy/sentinelone/network).

</details>

<details>

<summary>Does the EDR also monitor internal vulnerabilities?</summary>

No. Neither CrowdStrike nor SentinelOne detects software vulnerabilities in the standard EDR protection. It is an optional add-on that requires an additional subscription:

* CrowdStrike offers the Spotlight module.
* SentinelOne offers the Vulnerability Management module.

Without one of these modules, keep your own vulnerability monitoring, analysis and patching processes.

With one of these modules:

* Known vulnerabilities on endpoints are **detected automatically**, in near real-time, since the agent is deployed on the endpoints.
* The interface shows **which patches to apply**. You still apply the updates and patches yourself.
* Once the EDR is deployed, no further action is required: the Spotlight option is enabled automatically. In CrowdStrike, it is available to users with "Vulnerability management" rights.

</details>

<details>

<summary>How do I add more licences?</summary>

Deploy the EDR on new endpoints with your customer ID: you can add licences freely. An automatic annual adjustment appears on your invoice, based on the total number of activations.

</details>


# Setup: Microsoft 365

Connect your Microsoft 365 tenant so Email Security can analyze your mailboxes. Every mailbox gets the same features, whatever its Microsoft license.

## Prerequisites

* A Microsoft 365 **Global Administrator** account.

## Connect your tenant

{% stepper %}
{% step %}

### Sign in to Microsoft

1. In Stoïk Protect, go to `Email Security` > `Settings`.
2. Select `Microsoft 365` and click `Confirm`.
3. Sign in with your Global Administrator account.
   {% endstep %}

{% step %}

### Accept the permissions

Review and accept the requested permissions. See [Required permissions](https://docs.stoik.io/email-security/microsoft-365/permissions) for what each one is used for.
{% endstep %}

{% step %}

### Follow the guided setup

Stoïk Protect detects your license type and shows the right path:

* **No action needed**: you are redirected to the results page. The setup is complete.
* **Minimal action required**: enable audit in Microsoft Purview. See [Enable audit in Microsoft Purview](#enable-audit-in-microsoft-purview).
* **Full setup required**: configure Purview access first, then enable audit. See [Configure Microsoft Purview access](#configure-microsoft-purview-access).
  {% endstep %}
  {% endstepper %}

## Configure Microsoft Purview access

Only if Stoïk Protect asks for it and Purview access is not configured yet.

1. Go to the [Azure Portal](https://portal.azure.com/) and sign in with an administrator account.
2. Open the `Marketplace`.
3. Search for Microsoft Purview, select it, and click `Create`.

## Enable audit in Microsoft Purview

Only if Stoïk Protect asks for it.

1. Go to [Microsoft Purview](https://purview.microsoft.com/) and sign in with an administrator account.
2. Open `Solutions` > `Audit`.
3. Click `Start recording user and admin activity`.

If Purview shows an error, see [Purview activation errors](https://docs.stoik.io/email-security/microsoft-365/purview-errors).

## After the connection

* Monitored mailboxes appear in `Email Security` > `Mailboxes`. See [Mailbox statuses](https://docs.stoik.io/email-security/mailbox-statuses).
* Stoïk retrieves the emails of the past 90 days for all monitored mailboxes, starting with the oldest. This can take some time.

To connect another tenant (Microsoft 365 or Google Workspace), go to `Email Security` > `Settings`, click `Add a provider` and follow the same steps. You can connect as many tenants as you need.

{% hint style="info" %}
If the users list doesn't show up more than 10 minutes after synchronization, contact <protect@stoik.io>.
{% endhint %}


# Required permissions

The Microsoft permissions granted during the [setup](https://docs.stoik.io/email-security/microsoft-365). Stoïk uses them only to analyze emails and accounts, detect threats, and act quickly in the event of an incident.

| Permission                                                              | Usage                                                                                                                                                                                                         |
| ----------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Read all audit log data                                                 | Allows reading all audit logs across the organization for investigation and monitoring.                                                                                                                       |
| Read audit logs data from all services                                  | Allows the app to read audit logs from all Microsoft 365 services for threat detection and compliance.                                                                                                        |
| Read and write mail in all mailboxes                                    | Grants full read/write access to mail across every user mailbox (used for scanning, remediation, or actions on malicious emails).                                                                             |
| Read and write all user mailbox settings                                | Allows modifying mailbox settings (inbox rules, signatures, forwarding…) for all users.                                                                                                                       |
| Read your organization’s conditional access policies                    | Lets the app read Conditional Access configuration for audit or analysis of security posture.                                                                                                                 |
| Read all usage reports                                                  | Allows accessing organization-wide usage and activity reports (M365 usage analytics).                                                                                                                         |
| Read metadata and detection details for all emails in your organization | Allows the app to access email metadata (headers, detections, threat indicators) without accessing message bodies.                                                                                            |
| Read all users’ full profiles                                           | Grants full read access to users’ directory profiles (name, job info, attributes…).                                                                                                                           |
| Read and write all users’ authentication methods                        | Allows modifying users’ authentication methods (MFA settings, phone, email…).                                                                                                                                 |
| Sign in and read user profiles                                          | Allows the app to sign in as a user and access basic profile data.                                                                                                                                            |
| Read and write all password profiles and reset user passwords           | Allows resetting or updating passwords for all users. Required for post-incident remediation.                                                                                                                 |
| Revoke all sign in sessions for a user                                  | Allows the app to immediately revoke all active user sessions. Used for immediate containment.                                                                                                                |
| Read activity data for your organization                                | Allows Stoïk to read activity feed events in Microsoft 365 to detect suspicious actions or anomalies related to user activity.                                                                                |
| Read and write all applications                                         | Allows the app to read and write all applications and service principals in the directory. Used to analyze configuration, identify anomalies, and remove malicious applications used to maintain persistence. |
| Read and write all devices                                              | Allows the app to read and write device objects in the directory. Used to detect device-code phishing attacks and track compromised devices.                                                                  |


# Purview activation errors

Fix the errors Microsoft Purview can show after you enable audit logging during the [setup](https://docs.stoik.io/email-security/microsoft-365).

## "Trouble figuring out if activity is being recorded"

After enabling audit logging, you may see:

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2FhJA6xidfLH0Nn7Ap1YQ2%2Fpurview-error-message.png?alt=media&amp;token=75758ab0-84dd-443d-9a6f-bc4cb0048f41" alt=""><figcaption></figcaption></figure>

> Sorry, we're having trouble figuring out if activity is being recorded. Try refreshing the page.

In most cases this does **not** mean the activation failed. It reflects a delay on Microsoft's side between the moment audit logging is enabled and the moment it is effectively recording activity. This delay can take from a few minutes up to several hours, and it cannot be sped up from our side.

{% hint style="info" %}
The error message in the Purview web portal is not a reliable indicator. It does not indicate whether Purview is actually active, only that the portal cannot confirm it at that moment.
{% endhint %}

## "Enable-OrganizationCustomization" error

A different message means the tenant has never been customized (it is "dehydrated"). Audit logging cannot be turned on until that one-time setup has run:

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2Fgit-blob-45f07879a296d2f30f358312eef4e44101ae0de1%2Fpurview-dehydrated-error.png?alt=media" alt=""><figcaption></figcaption></figure>

> The command you tried to run isn't currently allowed in your organization. To run this command, you first need to run the command: Enable-OrganizationCustomization.

Unlike the delay above, this one requires an action. Two options:

* The portal shows a **Complete the organization configuration** prompt: clicking **Yes** runs the customization, but it does not always re-enable audit logging afterwards, so you often end up back on the error.
* The reliable fix is the PowerShell steps below, which run the customization and then re-enable audit logging in one go. The [Purview Script](https://docs.stoik.io/email-security/microsoft-365/purview-script) does this end to end: it runs the customization, waits for the tenant to finish hydrating, then enables audit logging automatically.

## Activate audit logging with PowerShell

The most reliable way to activate Purview audit logging is to force it directly through PowerShell, instead of relying on the web portal.

You can run these commands either from PowerShell installed on your machine, or from Azure Cloud Shell directly in the browser.

{% stepper %}
{% step %}

### Connect to Exchange Online

{% tabs %}
{% tab title="Local PowerShell" %}
Install the module, then connect. Replace `<admin@yourdomain.com>` with an account that has administrative permissions on the tenant.

```powershell
Install-Module -Name ExchangeOnlineManagement
Import-Module ExchangeOnlineManagement
Connect-ExchangeOnline -UserPrincipalName <admin@yourdomain.com>
```

{% endtab %}

{% tab title="Azure Cloud Shell" %}
Azure Cloud Shell requires an active Azure subscription on the tenant. If none is available, use local PowerShell instead.

Open [Azure Cloud Shell](https://shell.azure.com) in PowerShell mode. If the top of the terminal shows a `Switch to Bash` button, you are already in PowerShell mode.

Connect to Exchange Online:

```powershell
Connect-ExchangeOnline
```

<figure><img src="https://2582414397-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FaDGGQ49Oui29Ft0kJjSm%2Fuploads%2F6th9j4RFdqPwnXRYAn0W%2Fimage.png?alt=media&amp;token=6511635c-2dca-44b2-a424-1591f82f3247" alt=""><figcaption></figcaption></figure>
{% endtab %}
{% endtabs %}
{% endstep %}

{% step %}

### Force the activation

Run the commands one by one, in order. Forcing a disable then a re-enable is what reliably unblocks tenants stuck on the error.

1. Disable audit logging:

   ```powershell
   Set-AdminAuditLogConfig -UnifiedAuditLogIngestionEnabled $false
   ```
2. Re-enable it:

   ```powershell
   Set-AdminAuditLogConfig -UnifiedAuditLogIngestionEnabled $true
   ```
3. Confirm the status:

   ```powershell
   Get-AdminAuditLogConfig | Format-List UnifiedAuditLogIngestionEnabled
   ```

`UnifiedAuditLogIngestionEnabled` should now return `True`.

<details>

<summary>The re-enable command returns an error</summary>

On tenants that have never been customized, the re-enable command can fail with:

> The command you tried to run isn't currently allowed in your organization. To run this command, you first need to run the command: Enable-OrganizationCustomization.

Run the customization command, then check that the tenant is no longer dehydrated:

```powershell
Enable-OrganizationCustomization
Get-OrganizationConfig | Format-List IsDehydrated
```

`IsDehydrated` should return `False`. Once it does, run the re-enable command again.

</details>
{% endstep %}
{% endstepper %}

## After the command returns True

A `True` result means the command went through, but Purview is not active instantly: Microsoft takes time to actually start recording activity, sometimes several hours. Refresh the Purview portal after a while and the error message should be gone.

If the error is still there after several hours, run the force activation again (set `$false` then `$true`). If it still does not work after that, the issue is most likely tied to the tenant licensing or permissions. Contact our team at <protect@stoik.io> with the result of the command so we can help diagnose it.


# Purview script

The automated version of [Purview activation errors](https://docs.stoik.io/email-security/microsoft-365/purview-errors). It connects to Exchange Online, forces a disable then re-enable of audit log ingestion, and confirms the result. If the tenant is dehydrated, it runs `Enable-OrganizationCustomization`, waits for it to finish, then re-enables audit logging. If activation is genuinely refused, it prints a diagnostic (account roles, tenant hydration, license and subscription status).

## Prerequisites

* **Windows PowerShell** or **Azure Cloud Shell** (PowerShell mode). The script refuses to run on macOS/Linux, where the Exchange Online module fails on write operations.
* A tenant admin account.

{% hint style="info" %}
No Windows machine? If the tenant has an active Azure subscription, run it from [Azure Cloud Shell](https://shell.azure.com) in PowerShell mode: upload the script and run it without `-AdminUpn`. Cloud Shell reuses your session, so there is no sign-in prompt. Purview itself does not require an Azure subscription.
{% endhint %}

## Run the script

{% stepper %}
{% step %}

### Copy the script

Copy the [script](#script) with the copy button at the top right of the code block, and save it as `Repair-PurviewAuditLogging.ps1`.
{% endstep %}

{% step %}

### Allow script execution

In PowerShell, run this once per window (session-only, nothing permanent):

```powershell
Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass
```

{% endstep %}

{% step %}

### Run it with the admin account

Replace `<admin@clientdomain.com>` with a tenant admin. A browser sign-in window opens; the script does the rest.

```powershell
cd C:\path\to\the\folder
.\Repair-PurviewAuditLogging.ps1 -AdminUpn <admin@clientdomain.com>
```

{% endstep %}

{% step %}

### Verify later

`False` right after the run is **normal**: Microsoft applies it within a few minutes (up to about 60 min). Reconnect to check:

```powershell
Connect-ExchangeOnline -UserPrincipalName <admin@clientdomain.com> -ShowBanner:$false
Get-AdminAuditLogConfig | Format-List UnifiedAuditLogIngestionEnabled
```

`True` means it is done.
{% endstep %}
{% endstepper %}

## Troubleshooting

The script runs a diagnostic and tells you which case you are in. Send the `.log` file it created to <protect@stoik.io>.

* **Dehydrated tenant**: handled automatically. The script runs `Enable-OrganizationCustomization`, waits for hydration, then re-enables audit logging.
* **403 Forbidden / refused**: a real rejection. The account lacks the right role (the **Audit Logs** role, granted through Organization Management or Compliance Management), or the tenant has no license that includes Exchange Online.
* **Server-side error**: a transient "try again later" from Microsoft. The script already retries; if it persists, run it again a few minutes later.

## Script

{% code title="Repair-PurviewAuditLogging.ps1" lineNumbers="true" %}

```powershell
<#
.SYNOPSIS
    Repair Microsoft Purview unified audit logging by forcing a disable then re-enable.

.DESCRIPTION
    PREREQUISITE: run this on Windows PowerShell (5.1 or 7+) OR in Azure Cloud Shell.
    The script refuses to run anywhere else (macOS / desktop Linux) - those environments
    hit Exchange Online module quirks and are not supported.

      - Windows      -> pass -AdminUpn (or omit it for an interactive sign-in).
      - Cloud Shell  -> sign-in is automatic (the EXO module reuses your Cloud Shell identity).
                        Use -Device only if you specifically want device-code sign-in.

    Even when audit logging already reads as enabled, the Purview portal can stay stuck on
    its error. Forcing a disable then a re-enable is what unblocks it.

    Re-enabling is asynchronous: Microsoft can take up to ~60 minutes to reflect it, so a
    fresh read may still show False right after a successful call. That is normal.

    If enabling is genuinely refused, the script runs a diagnostic: connected account &
    roles, tenant hydration state, tenant age, and subscription/license status (the last two
    via Microsoft Graph when available) - to point at the real cause (permissions, licensing,
    or a too-new tenant).

    A timestamped .log file is written in the current folder.

.PARAMETER AdminUpn
    Admin UPN (e.g. admin@clientdomain.com). Optional on Windows (interactive sign-in if omitted).

.PARAMETER Device
    Force device-code sign-in. Optional - Cloud Shell signs in automatically without it.

.EXAMPLE
    .\Repair-PurviewAuditLogging.ps1 -AdminUpn admin@clientdomain.com
#>

[CmdletBinding()]
param(
    [string]$AdminUpn,
    [switch]$Device
)

$ErrorActionPreference = 'Stop'
$SupportEmail = 'protect@stoik.io'

# Patience for Microsoft's transient "try again later" errors.
$RetryCount   = 8
$RetryWaitSec = 30   # 8 x 30s = up to ~4 min per toggle

# --- Console helpers --------------------------------------------------------
function Write-Step { param($m) Write-Host "`n==> $m" -ForegroundColor Cyan }
function Write-Ok   { param($m) Write-Host "    [OK]   $m" -ForegroundColor Green }
function Write-Info { param($m) Write-Host "    [..]   $m" -ForegroundColor Gray }
function Write-Warn2{ param($m) Write-Host "    [WARN] $m" -ForegroundColor Yellow }
function Write-Err  { param($m) Write-Host "    [FAIL] $m" -ForegroundColor Red }

# --- Environment detection + prerequisite gate ------------------------------
# Cloud Shell sets one of these; Windows PowerShell 5.1 has no $IsWindows so we also
# accept the Desktop edition.
$IsCloudShell = [bool]$env:ACC_CLOUD `
                -or ($env:POWERSHELL_DISTRIBUTION_CHANNEL -like '*CloudShell*') `
                -or ($env:AZUREPS_HOST_ENVIRONMENT -like '*cloud-shell*')
$OnWindows    = $IsWindows -or ($PSVersionTable.PSEdition -eq 'Desktop')

$exoVersion = (Get-Module -ListAvailable -Name ExchangeOnlineManagement |
               Sort-Object Version -Descending | Select-Object -First 1).Version

Write-Step "Environment"
Write-Host "    OS / platform : $([System.Environment]::OSVersion.VersionString)"
Write-Host "    PowerShell    : $($PSVersionTable.PSVersion) ($($PSVersionTable.PSEdition))"
Write-Host "    Context       : $(if ($IsCloudShell) {'Azure Cloud Shell'} elseif ($OnWindows) {'Windows'} else {'unsupported'})"
Write-Host "    EXO module    : $(if ($exoVersion) {$exoVersion} else {'not installed'})"

if (-not ($OnWindows -or $IsCloudShell)) {
    Write-Err "Unsupported environment."
    Write-Host "    This script only runs on Windows PowerShell or in Azure Cloud Shell." -ForegroundColor Yellow
    Write-Host "    Re-run it from a Windows machine, or from https://shell.azure.com (PowerShell)." -ForegroundColor Yellow
    exit 2
}

# --- Logging ----------------------------------------------------------------
$timestamp = Get-Date -Format 'yyyyMMdd-HHmmss'
$LogFile   = Join-Path (Get-Location) "purview-repair-$timestamp.log"
Start-Transcript -Path $LogFile -Append | Out-Null

# Write text to a side file without ever throwing (the .log is held by the transcript).
function Save-ErrorDetail {
    param($Text)
    if ([string]::IsNullOrWhiteSpace($Text)) { return $null }
    try {
        $f = [System.IO.Path]::ChangeExtension($LogFile, 'error.txt')
        "FULL ERROR DETAIL (for support)`n$Text" | Out-File -FilePath $f -ErrorAction Stop
        return $f
    } catch { return $null }
}

$connected = $false

# Runs Set-AdminAuditLogConfig once and classifies the outcome. EXO REST cmdlets don't
# always honour try/catch, so we inspect -ErrorVariable.
function Set-AuditIngestionOnce {
    param([bool]$Value)
    $ErrorActionPreference = 'Continue'
    # Hide the raw "...up to 60 minutes to apply" warning from the console (we still capture
    # it via -WarningVariable and print our own [OK]). This is a preference, not a command
    # decoration or a stream redirect.
    $WarningPreference = 'SilentlyContinue'
    $err = $null; $warn = $null
    # IMPORTANT: invoke this EXACTLY like a hand-typed command. Decorating the call with
    # -WarningAction SilentlyContinue / 2>$null / | Out-Null makes EXO return a spurious (403)
    # inside a script, while the bare command succeeds. -ErrorVariable / -WarningVariable only
    # capture the streams; they do not change how the cmdlet talks to the server.
    $null = Set-AdminAuditLogConfig -UnifiedAuditLogIngestionEnabled $Value `
        -ErrorVariable err -WarningVariable warn
    $errText  = ($err  | ForEach-Object { $_.ToString() }) -join ' '
    $warnText = ($warn | ForEach-Object { $_.ToString() }) -join ' '
    if ($errText) { $script:LastFullError = $errText }

    $dehydrated = ($errText -match 'Enable-OrganizationCustomization')
    $serverSide = ($errText -match 'server side error|try again after some time')
    $forbidden  = (-not $serverSide) -and ($errText -match '403|Forbidden|Interdit|Access is denied|not authorized|insufficient')

    if     ($dehydrated) { $short = 'Tenant is dehydrated (needs Enable-OrganizationCustomization).' }
    elseif ($forbidden)  { $short = '403 Forbidden - permissions / role / licensing problem.' }
    elseif ($serverSide) { $short = "Microsoft is busy (server-side error, 'try again later')." }
    elseif ($errText)    { $short = ($errText -split "`n")[0] }
    else                 { $short = '' }

    [pscustomobject]@{
        Ok = [string]::IsNullOrWhiteSpace($errText); Warning = $warnText; Error = $errText
        Short = $short; Dehydrated = $dehydrated; Forbidden = $forbidden; ServerSide = $serverSide
    }
}

# Toggle the setting, retrying on transient errors only.
function Invoke-AuditToggle {
    param([bool]$Value, [string]$Label)
    $r = $null
    for ($i = 1; $i -le $RetryCount; $i++) {
        $r = Set-AuditIngestionOnce $Value
        if ($r.Ok)         { Write-Ok "$Label accepted by Microsoft."; return $r }
        if ($r.Dehydrated) { return $r }                       # caller hydrates then retries
        if ($r.Forbidden)  { Write-Warn2 "$Label refused (permissions / role / licensing) - not retrying."; return $r }

        Write-Warn2 "$Label failed (attempt $i/$RetryCount): $($r.Short)"
        if ($i -lt $RetryCount) {
            Write-Info "Retrying $Label in ${RetryWaitSec}s."
            Start-Sleep -Seconds $RetryWaitSec
        }
    }
    return $r
}

# Manual checklist shown when Graph can't auto-check subscription / licensing / tenant age.
function Write-ManualChecklist {
    Write-Info "Verify manually in the Microsoft 365 admin center (admin.microsoft.com):"
    Write-Info "  - Billing > Your products : the subscription is Active (not suspended/expired)."
    Write-Info "  - A license that includes Exchange Online is assigned (most M365/O365 plans)."
    Write-Info "  - Setup date : tenants created in the last ~72h may not have audit provisioned yet."
}

# Full diagnostic when enabling is genuinely refused.
function Invoke-FailureDiagnostic {
    param($ErrInfo)
    Write-Step "Diagnostic - why was enabling refused?"
    if ($ErrInfo -and $ErrInfo.Short) { Write-Info "Reported error : $($ErrInfo.Short)" }

    # Connected identity
    $upn = $null
    try {
        $ci  = Get-ConnectionInformation -ErrorAction Stop | Select-Object -First 1
        $upn = $ci.UserPrincipalName
        Write-Info "Connected as   : $upn"
        Write-Info "Tenant id      : $($ci.TenantId)"
    } catch { Write-Info "Could not read connection info." }

    # Tenant / org configuration
    try {
        $org = Get-OrganizationConfig -ErrorAction Stop
        Write-Info "Organization   : $($org.DisplayName) [$($org.Name)]"
        Write-Info "IsDehydrated   : $($org.IsDehydrated)$(if ($org.IsDehydrated) {'  <- run Enable-OrganizationCustomization'})"
    } catch { Write-Info "Could not read Get-OrganizationConfig." }

    # Permissions: does the account hold a role that can set audit config?
    if ($upn) {
        try {
            $roles = Get-ManagementRoleAssignment -RoleAssignee $upn -Delegating:$false -ErrorAction Stop |
                     Select-Object -ExpandProperty Role -Unique
            Write-Info "Roles          : $([string]::Join(', ', $roles))"
            if ($roles -contains 'Audit Logs' -or $roles -contains 'Organization Configuration') {
                Write-Ok "Account holds a role allowing audit configuration."
            } else {
                Write-Warn2 "Account lacks 'Organization Configuration' / 'Audit Logs'."
                Write-Warn2 "=> Add it to the 'Organization Management' role group (Exchange admin center)."
            }
        } catch { Write-Info "Could not enumerate roles ($($_.Exception.Message))." }
    }

    # Subscription / licensing / tenant age via Microsoft Graph (best-effort, non-fatal).
    if (Get-Command Get-MgOrganization -ErrorAction SilentlyContinue) {
        try {
            if (-not (Get-MgContext -ErrorAction SilentlyContinue)) {
                Write-Info "Connecting to Microsoft Graph (read-only) for subscription/license/age checks..."
                Connect-MgGraph -Scopes 'Organization.Read.All' -NoWelcome -ErrorAction Stop
            }
            $o = Get-MgOrganization -ErrorAction Stop | Select-Object -First 1
            if ($o.CreatedDateTime) {
                $age = (New-TimeSpan -Start $o.CreatedDateTime -End (Get-Date)).Days
                Write-Info "Tenant created : $($o.CreatedDateTime) (~$age days ago)"
                if ($age -lt 3) { Write-Warn2 "Tenant is very new (<3 days) - audit provisioning may not be ready. Retry later." }
            }
            foreach ($s in (Get-MgSubscribedSku -ErrorAction Stop)) {
                Write-Info "License        : $($s.SkuPartNumber)  status=$($s.CapabilityStatus)  seats=$($s.PrepaidUnits.Enabled)"
                if ($s.CapabilityStatus -in 'Suspended','Deleted') {
                    Write-Warn2 "Subscription '$($s.SkuPartNumber)' is $($s.CapabilityStatus) - not active."
                }
            }
        } catch {
            Write-Info "Graph check unavailable: $($_.Exception.Message)"
            Write-ManualChecklist
        }
    } else {
        Write-Info "Microsoft Graph module not installed - cannot auto-check subscription/license/tenant age."
        Write-ManualChecklist
    }
}

try {
    # --- Prerequisites -----------------------------------------------------
    Write-Step "Checking prerequisites"
    if (-not $exoVersion) {
        Write-Warn2 "ExchangeOnlineManagement not found. Installing for the current user."
        if ($OnWindows -and $PSVersionTable.PSEdition -eq 'Desktop') {
            # Windows PowerShell 5.1 often needs TLS 1.2 explicitly to reach the PSGallery.
            [Net.ServicePointManager]::SecurityProtocol = [Net.SecurityProtocolType]::Tls12
        }
        Install-Module -Name ExchangeOnlineManagement -Scope CurrentUser -Force -AllowClobber
        Write-Ok "Module installed."
    } else {
        Write-Ok "ExchangeOnlineManagement present ($exoVersion)."
    }
    Import-Module ExchangeOnlineManagement

    # --- Connect -----------------------------------------------------------
    Write-Step "Connecting to Exchange Online"
    # In Azure Cloud Shell, the EXO module reuses the session identity automatically - NO device
    # code. Only force device-code when the user explicitly asks for it with -Device.
    $useDevice = [bool]$Device
    if ($useDevice -and $AdminUpn) {
        Write-Warn2 "Device-code sign-in: open the URL shown below and enter the code."
        Connect-ExchangeOnline -UserPrincipalName $AdminUpn -Device -ShowBanner:$false
    } elseif ($useDevice) {
        Write-Warn2 "Device-code sign-in: open the URL shown below and enter the code."
        Connect-ExchangeOnline -Device -ShowBanner:$false
    } elseif ($AdminUpn) {
        Connect-ExchangeOnline -UserPrincipalName $AdminUpn -ShowBanner:$false
    } else {
        Connect-ExchangeOnline -ShowBanner:$false
    }
    $connected = $true
    Write-Ok "Connected."

    # --- Diagnostic --------------------------------------------------------
    Write-Step "Current state"
    $enabled      = (Get-AdminAuditLogConfig).UnifiedAuditLogIngestionEnabled
    $isDehydrated = (Get-OrganizationConfig).IsDehydrated
    Write-Host "    UnifiedAuditLogIngestionEnabled : $enabled"
    Write-Host "    IsDehydrated                    : $isDehydrated"

    # --- Disable first only if currently ON (the off->on cycle unblocks the portal) -------
    if ($enabled) {
        Write-Step "Step 1 - Disabling ingestion (to force a clean off->on cycle)"
        $d = Invoke-AuditToggle -Value $false -Label "Disable"
        if (-not $d.Ok) { Write-Warn2 "Disable did not confirm ($($d.Short)) - continuing to re-enable anyway." }
    } else {
        Write-Info "Audit logging is currently OFF - enabling directly (no disable needed)."
    }

    # --- Enable ------------------------------------------------------------
    Write-Step "Step 2 - Enabling ingestion"
    $e = Invoke-AuditToggle -Value $true -Label "Enable"
    if ($e.Dehydrated) {
        Write-Warn2 "Dehydrated tenant. Running Enable-OrganizationCustomization."
        try { Enable-OrganizationCustomization } catch { Write-Warn2 $_.Exception.Message }
        $h = 0
        do {
            Start-Sleep -Seconds 15
            $isDehydrated = (Get-OrganizationConfig).IsDehydrated
            $h++
            Write-Info "Waiting for hydration... attempt $h (IsDehydrated = $isDehydrated)"
        } while ($isDehydrated -and $h -lt 8)
        if ($isDehydrated) { throw "Tenant is still dehydrated after waiting. Re-run in a few minutes." }
        Write-Ok "Tenant hydrated."
        $e = Invoke-AuditToggle -Value $true -Label "Re-enable"
    }

    # --- Result ------------------------------------------------------------
    Write-Step "Result"
    Start-Sleep -Seconds 5
    $final = (Get-AdminAuditLogConfig).UnifiedAuditLogIngestionEnabled
    Write-Host "    UnifiedAuditLogIngestionEnabled (fresh read) : $final"

    # The fresh read is the ground truth: if it is True, audit logging IS enabled, even when the
    # last enable call left $e.Ok false (transient 403, "try again" noise on the error stream).
    if ($final) {
        Write-Ok "Audit logging is ENABLED. Done."
    }
    elseif ($e.Ok) {
        Write-Warn2 "Re-enable was accepted but still reads False - normal: Microsoft can take up to"
        Write-Host  "    ~60 minutes to apply it. Re-check with the command below shortly." -ForegroundColor Yellow
    }
    else {
        Write-Err "Enabling was NOT accepted by Microsoft. Running a diagnostic."
        Save-ErrorDetail $script:LastFullError | Out-Null
        Invoke-FailureDiagnostic $e
    }

    # --- Verify later ------------------------------------------------------
    Write-Step "Verify later (reconnect first, the script disconnects at the end)"
    if ($AdminUpn) {
        Write-Host  "        Connect-ExchangeOnline -UserPrincipalName $AdminUpn -ShowBanner:`$false" -ForegroundColor Gray
    } else {
        Write-Host  "        Connect-ExchangeOnline -ShowBanner:`$false" -ForegroundColor Gray
    }
    Write-Host  "        Get-AdminAuditLogConfig | Format-List UnifiedAuditLogIngestionEnabled" -ForegroundColor Gray
    Write-Host  "    'True' means it is enabled."
}
catch {
    Write-Err "Something went wrong: $($_.Exception.Message)"
    if ($connected) { try { Invoke-FailureDiagnostic $null } catch { } }
    $errorFile = Save-ErrorDetail $script:LastFullError
    Write-Host ""
    Write-Host "    >> If you need help, send the log file to $SupportEmail :" -ForegroundColor Yellow
    Write-Host "       $LogFile" -ForegroundColor Yellow
    if ($errorFile) { Write-Host "       $errorFile" -ForegroundColor Yellow }
}
finally {
    if ($connected) {
        Write-Step "Disconnecting"
        try { Disconnect-ExchangeOnline -Confirm:$false } catch { }
        Write-Ok "Disconnected."
    }
    Write-Host "`nLog saved to: $LogFile" -ForegroundColor Gray
    Stop-Transcript | Out-Null
}
```

{% endcode %}


# Setup: Google Workspace

Connect your Google Workspace tenant so Email Security can analyze your mailboxes.

## Prerequisites

* A **Google Workspace Admin** account.
* Google Chrome. Safari cannot be used.
* Stoïk Protect and the Google Admin Console open in two tabs of the **same browser window**.

## Connect your tenant

{% stepper %}
{% step %}

### Sign in with your admin account

1. In Stoïk Protect, go to `Email Security`.
2. Select Google Workspace and click `Connect your mailbox`.
3. Sign in with your Google Workspace Admin account.

{% hint style="info" %}
If you get an `Access blocked` / `Error 400: access_not_configured` screen instead of the Google sign-in prompt, see [Error 400: access\_not\_configured](https://docs.stoik.io/human-security/phishing/setup-errors#error-400-access_not_configured) for the fix. The same restrictive app-authorization policy blocks both the Phishing module and Email Security.
{% endhint %}
{% endstep %}

{% step %}

### Authorize Stoïk in the Google Admin Console

1. Open [Domain-wide delegation](https://admin.google.com/ac/owl/domainwidedelegation) in the Google Admin Console.
2. Click `Add new`.
3. Enter the Client ID `108893255833546679906` and the OAuth scopes below. See [Required permissions](https://docs.stoik.io/email-security/google-workspace/permissions) for what each one is used for.

   ```
   https://www.googleapis.com/auth/gmail.readonly,https://www.googleapis.com/auth/gmail.modify,https://www.googleapis.com/auth/gmail.insert,https://www.googleapis.com/auth/gmail.metadata,https://www.googleapis.com/auth/gmail.settings.basic,https://www.googleapis.com/auth/admin.directory.user.readonly,https://www.googleapis.com/auth/admin.reports.audit.readonly,https://www.googleapis.com/auth/apps.alerts,https://www.googleapis.com/auth/admin.directory.customer.readonly,https://www.googleapis.com/auth/gmail.settings.sharing,https://www.googleapis.com/auth/admin.directory.device.mobile.readonly
   ```
4. Click `Authorize`.
   {% endstep %}

{% step %}

### Confirm

In Stoïk Protect, click `Confirm` to verify the configuration. Your Google tenant is connected.
{% endstep %}
{% endstepper %}

## After the connection

* Monitored mailboxes appear in `Email Security` > `Mailboxes`. See [Mailbox statuses](https://docs.stoik.io/email-security/mailbox-statuses).
* Stoïk retrieves the emails of the past 90 days for all monitored mailboxes, starting with the oldest. This can take some time.

To connect another tenant (Microsoft 365 or Google Workspace), go to `Email Security` > `Settings`, click `Add a provider` and follow the same steps. You can connect as many tenants as you need.

{% hint style="info" %}
If the users list doesn't show up more than 10 minutes after synchronization, contact <protect@stoik.io>.
{% endhint %}


# Required permissions

The OAuth scopes authorized in the Google Admin Console during the [setup](https://docs.stoik.io/email-security/google-workspace). Stoïk uses them only to analyze emails and accounts, detect threats, and act quickly in the event of an incident.

| Permission                               | Usage                                                                                            |
| ---------------------------------------- | ------------------------------------------------------------------------------------------------ |
| `gmail.readonly`                         | Read messages and settings in the mailboxes, to analyze emails.                                  |
| `gmail.modify`                           | Change messages and labels in the mailboxes (no permanent deletion), to act on malicious emails. |
| `gmail.insert`                           | Insert or import a message into a mailbox only.                                                  |
| `gmail.metadata`                         | Read headers and labels only, with no body and no attachments.                                   |
| `gmail.settings.basic`                   | Read and change filters, vacation reply, IMAP, and POP settings.                                 |
| `gmail.settings.sharing`                 | Manage delegates and mail-forwarding addresses.                                                  |
| `admin.directory.user.readonly`          | List users and read their directory profiles.                                                    |
| `admin.directory.customer.readonly`      | Read the Workspace customer record (organization name, primary domain).                          |
| `admin.directory.device.mobile.readonly` | List managed mobile devices and their metadata.                                                  |
| `admin.reports.audit.readonly`           | Read the Admin audit logs (logins, admin actions, Gmail activity).                               |
| `apps.alerts`                            | Read, delete, and give feedback on Alert Center alerts.                                          |


# Mailbox statuses

All the mailboxes monitored by Email Security are listed in `Email Security` > `Mailboxes`, with their protection status.

## Mailbox details

* **User**: the associated account owner.
* **Email address**: the exact monitored address.
* **Tenant**: the tenant linked to the mailbox.
* **Type**: `User` (personal mailbox, assigned to one person) or `Shared` (mailbox used by several people, with no single owner).
* **Suspicious emails**: the status of email analysis.
* **Compromised accounts**: the status of account takeover detection.

## Suspicious emails status

* **Active**: incoming and outgoing emails are being analyzed.
* **Inactive**: emails on this mailbox are not being analyzed. Hover over the status to see the reason, for example `Connection expired`, `Access revoked` or `Temporary sync issue`.

## Compromised accounts status

* **Active**: enough sign-in history to detect unusual account behavior.
* **Analyzing**: still gathering sign-in history. Detection is on, but not at full precision yet.
* **Dormant**: too little recent activity, so detection precision is reduced.

## Mailbox type

* **Microsoft 365**: pulled directly from the mailbox's configuration in Outlook (user or shared). If a mailbox shows an unexpected type, check its Outlook configuration.
* **Google Workspace**: a mailbox showing as `Shared` is usually a "Group" with individual mailboxes assigned to it, not an Outlook-style shared mailbox.

{% hint style="info" %}
For Google tenants, the user or shared type is only available from a recent change onwards: older data is not filled in. Microsoft 365 tenants have full historical data.
{% endhint %}


# Suspicious emails

Email Security detects suspicious or malicious emails in your mailboxes. The Stoïk SOC reviews every alert: no action is required on your side.

There are two types of email threats:

* **Fraud:** financial fraud attempts, such as invoice fraud or impersonation of a known contact.
* **Phishing:** emails linking to a page that impersonates a known brand to steal credentials.

## Alerts table

Alerts are listed in `Email Security` > `Suspicious emails`. Each row shows:

* **Subject:** the email subject line
* **Criticality:** Low, Medium, High, or Critical
* **Protected user:** the mailbox that received or sent the email
* **Detection date:** when the email was received
* **Status:** Open or Closed
* **Validity:** Undefined, False Positive or True Positive
* **Actions:** None, Banner and/or Quarantine

Click an alert to open its details.

## Alert details

### Information

* **Detection date**: when the email was received
* **Protected user**: the mailbox that received or sent the email
* **Traffic**: inbound or outbound
* **Mailbox**: Microsoft 365 or Google Workspace
* **Type**: Fraud or Phishing
* **Reasons**: why the email was flagged (e.g. Impersonating user, Suspicious banking info)
* **Actions**: the actions taken by the SOC (e.g. email moved to quarantine, warning banner added). See [Remediation actions](https://docs.stoik.io/email-security/remediation-actions).
* **Status**: Open or Closed
* **Validity**: True Positive or False Positive

{% hint style="info" %}
Highly suspicious emails are not deleted. They are moved to a `Stoïk Quarantine` folder, created automatically on the mailbox at the first quarantine action, so they remain recoverable.
{% endhint %}

### Verdict

A colored banner at the top summarizes the assessment. Its color and headline match the alert's criticality and type. For a fraud alert:

<table><thead><tr><th width="173.8203125">Criticality</th><th>Headline</th></tr></thead><tbody><tr><td>Critical</td><td>Critical fraud attempt detected</td></tr><tr><td>High</td><td>High-risk fraud attempt detected</td></tr><tr><td>Medium</td><td>Suspicious activity detected</td></tr><tr><td>Low</td><td>Minor anomalies detected</td></tr></tbody></table>

For a phishing alert, the headline reads "phishing" instead of "fraud" (for example, "Critical phishing attempt detected"). A short summary below the headline gives more context.

### Attacker

Closed by default. It shows the sender name, sender domain and sender email address.

### Email content

The original email, as it landed in the inbox: senders, recipients, subject, sending date, the warning banner added by the SOC if any, and the full body with its formatting, images and links. Click **Expand** to read it in a larger window.

Attachments are listed below. Stoïk analyzes attachments (such as PDFs) for phishing and fraud indicators, for example embedded links or IBANs, but does not scan them for malware. Protection against malicious files relies on the security solution installed on your devices.

## Activity timeline

The Activity section lists everything that happened to the alert, most recent first.

### Alert handling

<table><thead><tr><th width="210.82421875">Step</th><th>Meaning</th></tr></thead><tbody><tr><td>Alert creation</td><td>When the email was originally received.</td></tr><tr><td>Alert confirmation</td><td>When the alert was ingested and confirmed in the Stoik system.</td></tr><tr><td>Alert processing</td><td>When a SOC analyst started investigating the alert.</td></tr><tr><td>Alert closure</td><td>When the alert was resolved and closed.</td></tr></tbody></table>

**Alert creation** and **Alert confirmation** always appear. **Alert processing** and **Alert closure** appear as the alert progresses.

{% hint style="info" %}
When Email Security is first activated, emails received up to 90 days before activation are scanned. For these emails, the gap between **Alert creation** and **Alert confirmation** can be large. This is expected.
{% endhint %}

### Signal checks

Each check shows a colored dot: a problem was found, an anomaly was detected, or the check passed.

<table><thead><tr><th width="239.77734375">Check</th><th>What it does</th></tr></thead><tbody><tr><td>Sender identity</td><td>Verifies if the sender is impersonating a known contact in your organization.</td></tr><tr><td>Communication history</td><td>Verifies if this sender has previously communicated with your organization.</td></tr><tr><td>Sender reputation</td><td>Checks the sender's domain and email address against known threat databases.</td></tr></tbody></table>

The description of each check reflects its actual result. For example, `Sender reputation` reads "The sender's domain or email address matches known threat databases." when flagged, and "The sender's domain and email address were not found in known threat databases." when passed.

### SOC actions, notes and emails

* **SOC actions and updates**: warning banners, moves to quarantine, and validity changes (for example from **Undefined** to **True Positive**).
* **Analyst notes**: context written by the Stoïk SOC on how the alert was investigated, when relevant.
* **Email exchanges**: the emails exchanged between Stoïk and your organization about the alert.


# Email search

Sometimes you need to look for an email yourself: a message an employee reported, a sender you are wary of, or a mail you want to trace across mailboxes. You can search your employees' mailboxes directly from Stoïk Protect, with no need to contact us.

{% stepper %}
{% step %}

## Open the search

In Stoïk Protect, go to `Email Security` > `Suspicious emails`, then open the **Search an email** tab.
{% endstep %}

{% step %}

## Enter your criteria

You can combine several of them:

* the email address of the sender or the recipient;
* the display name of the sender;
* the subject;
* a timeframe: the last 7 days, the last month, or a custom period.
  {% endstep %}

{% step %}

## Launch the search

Start the search and follow the results as they appear.
{% endstep %}
{% endstepper %}

## Read the results

* Results appear as the search runs. The first ones show up as soon as they are found, so you do not have to wait for the full search to finish.
* Click a result to open a preview of the email's content.
* The search keeps running in the background. You can switch tabs and come back to it later.

## Permission

Access to the search is controlled by a dedicated permission: **Search for an email**, under **Cyber: Managed Services**. It is unchecked by default, because reading employees' emails is more sensitive than reviewing suspicious alerts. Only people who are explicitly granted this permission can read emails.

{% hint style="info" %}
If you cannot see the tab or run a search, ask the person who manages permissions on your account to grant it to you.
{% endhint %}


# Compromised accounts

Email Security detects mailboxes showing signs of unauthorized access or suspicious activity. The Stoïk SOC reviews every alert and contains the account when needed.

## Alerts table

Alerts are listed in `Email Security` > `Compromised accounts`. The **Detected signals** column lists the signals behind each alert, grouped by type with their number of occurrences (e.g. `3 Impossible travels · 2 Abnormal behaviors`).

Click an alert to open its details.

## Alert details

### Information

* **Detection date:** when the compromise was detected
* **Status:** Open or Closed
* **Validity:** True Positive or False Positive
* **Protected user:** the mailbox that was targeted
* **Actions:** None, MFA reset, Password reset, etc. See [Remediation actions](https://docs.stoik.io/email-security/remediation-actions#compromised-accounts).
* **Criticality:** Low, Medium, High, or Critical

### User profile

A baseline of what's normal for this user, computed from the last 13 weeks of sign-in activity. Use it as a reference when reading the timeline.

<table><thead><tr><th width="235.86328125">Row</th><th>What it tells you</th></tr></thead><tbody><tr><td>Usual location</td><td>The top countries this user signs in from (with the share of sign-ins for each).</td></tr><tr><td>Usual device</td><td>The operating system and browser the user typically uses.</td></tr><tr><td>Usual sign-in pattern</td><td>A weekly heatmap showing when the user is normally active.</td></tr><tr><td>VPN usage</td><td>How often this user signs in through a VPN.</td></tr><tr><td>Email throughput</td><td>How many emails this user sends per day on average, with their historical peak.</td></tr><tr><td>Coverage</td><td>The number of sign-ins the baseline is computed from.</td></tr></tbody></table>

For example, if the user normally signs in from Paris on weekdays between 9 AM and 6 PM, a sign-in from Moscow at 3 AM is a strong anomaly.

## Activity timeline

The signals that triggered the alert and the actions taken by the Stoïk SOC, grouped by day, most recent first.

### Detected signals

* Impossible travel detected: two sign-ins too far apart given the time between them.
* Sign-in from high-risk country: sign-in from a country flagged in our threat database.
* Sign-in from unusual location / network / device / browser: the user signed in from a context they don't usually use.
* Sign-in at unusual time: sign-in outside of the user's normal working hours.
* Automated access detected: sign-in coming from a headless browser or automation tool.
* Suspicious app access: access from a tool often used by attackers (e.g. Graph Explorer, Azure CLI).
* High email volume detected: the user sent significantly more emails than usual in 24 hours.
* Suspicious inbox rule created: a rule was created on the mailbox that auto-deletes, forwards or hides messages.
* Authentication method added or removed: an MFA method was changed on the account.
* Sign-in from known malicious IP: sign-in from an IP listed in our threat feed.
* VPN used: sign-in via VPN, on a user who rarely uses one.
* Session reused across devices: the same session ID was seen from two different OS/browser combinations.
* Microsoft Entra flagged high risk: Microsoft's own identity protection raised an alert.
* Failed MFA attempt: an MFA challenge failed.

### Alert handling

<table><thead><tr><th width="235.421875">Step</th><th>Meaning</th></tr></thead><tbody><tr><td>Alert creation</td><td>When the compromise was detected.</td></tr><tr><td>Alert confirmation</td><td>When the alert was ingested and confirmed in the Stoik system.</td></tr><tr><td>Alert processing</td><td>When a SOC analyst started investigating the alert.</td></tr><tr><td>Alert closure</td><td>When the alert was resolved and closed.</td></tr></tbody></table>

**Alert creation** and **Alert confirmation** always appear. **Alert processing** and **Alert closure** appear as the alert progresses.

### SOC actions, notes and emails

* **SOC actions**: the [remediation actions](https://docs.stoik.io/email-security/remediation-actions#compromised-accounts) taken to contain the account.
* **Analyst notes**: context written by the Stoïk SOC on how the alert was investigated, when relevant.
* **Email exchanges**: the emails exchanged with your organization about the alert, when there are any.


# Remediation actions

The actions the Stoïk SOC can take for each type of Email Security detection.

## Suspicious emails

### Warning banner

A warning banner is added to the email. Its wording depends on the type of threat detected.

<details>

<summary>Phishing</summary>

* **Suspected Impersonation**: This email may impersonate a well-known brand (Microsoft, DocuSign, your bank, etc.) in order to steal your credentials. Do not click any links and do not enter your credentials. If in doubt, go directly to the official website by typing its address into your browser.
* **Suspected Quishing**: This email contains a QR code that may redirect you to a phishing page designed to steal your credentials or personal information. Do not scan this QR code with your phone. If in doubt, go directly to the website by typing its address into your browser and verify the request through a trusted channel.
* **Suspected Phishing via Attachment**: This email may contain an attachment (PDF, HTML, etc.) concealing a phishing link designed to steal your credentials or personal information. Do not open the attachment and do not click any links it contains. If in doubt, verify the request through a trusted channel.
* **Suspected Phishing via Link**: This email contains a link to a legitimate file-sharing service (SharePoint, OneDrive, Google Drive, Dropbox) that may be used as a gateway to a phishing page designed to steal your credentials. Do not enter your credentials on the page that opens. If in doubt, verify with the sender through a trusted channel before opening the file.
* **Suspected Spear Phishing**: This email may be a highly targeted phishing attempt, personalized with details about you, your role, or your organization to make it more convincing. Do not click any links and do not download any attachments. Verify the sender's identity through a trusted channel before replying.
* **Suspected Phishing** (fallback, used when no specific type of phishing is detected): This email may contain phishing links: URLs designed to steal your credentials or personal information. If in doubt, go directly to the website by typing its address into your browser and verify the request through a trusted channel.

</details>

<details>

<summary>Fraud</summary>

* **Suspected Fraudulent Contact**: This email may be an initial contact aimed at setting up financial fraud (for example, a request to change an IBAN or pay an invoice). If in doubt, please carry out a double check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.
* **Suspected Sextortion**: This email may be an extortion attempt threatening to disclose intimate images or videos of you unless a payment is made. This is a common scam, and no such content typically exists. Do not pay, do not reply, and report this email to your IT manager.
* **Suspected Extortion**: This email may be an extortion attempt: a threat to exploit an alleged vulnerability or supposedly stolen data unless a payment is made. Do not pay, do not reply, and report this email to your IT manager.
* **Suspected Gift Card Scam**: This email may be a scam asking you to purchase gift cards and send the codes to the sender, often under the pretext of an urgent request from a colleague or acquaintance. If in doubt, please double-check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.
* **Suspected CEO Fraud**: This email may be an attempt to impersonate an executive in order to push you into carrying out an urgent financial action (wire transfer, gift cards, change of bank details) or into disclosing confidential information. If in doubt, please carry out a double check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.
* **Suspected Vendor Fraud**: This email may impersonate a known supplier or business partner in order to push a fraudulent invoice or a change of bank details. If in doubt, please carry out a double check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.
* **Suspected Payroll Diversion**: This email may be an attempt to redirect an employee's salary payment to a fraudulent bank account. If in doubt, please carry out a double check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.
* **Suspected Fraudulent Request**: This email may contain a request to change a bank account or IBAN for a supplier or a third-party payment. If in doubt, please carry out a double check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.
* **Suspected Financial Fraud** (fallback, used when no specific type of fraud is detected): This email may be a fraud attempt involving a financial request or an invoice. If in doubt, please carry out a double check before taking any action: contact the person only through their usual phone number, never through the contact details mentioned in this message.

</details>

<details>

<summary>Malware</summary>

* **Suspected Malicious Attachment**: This email may contain a malicious attachment (document with macro, executable, booby-trapped archive) intended to infect your device. Do not open or download this attachment. If in doubt, check with the sender through a trusted channel before opening it.
* **Suspected Malicious Link**: This email may contain a link leading to a malicious payload intended to infect your device. If in doubt, verify the link's destination or contact the sender through a trusted channel before clicking.
* **Suspected Fake CAPTCHA**: This email or page may redirect you to a fake CAPTCHA or "human verification" screen that prompts you to run a command or script on your computer. No legitimate website asks you to run a command to prove you are human. If in doubt, close the page immediately and report it to your IT manager team.
* **Suspected Malicious Content** (fallback, used when no specific type of malware is detected): This email may contain malicious content (link, attachment, or embedded code) intended to infect your device or compromise your data. Do not click any links and do not open any attachments. If in doubt, check with the sender through a trusted channel before taking any action.

</details>

<details>

<summary>Spam</summary>

* **Suspicion of Unsolicited Prospecting**: This email may be unsolicited commercial prospecting: an offer, an introduction, or a proposal you did not request. Do not reply and do not click any link if you did not initiate this exchange. If in doubt, ignore this message and verify the sender's identity through a trusted channel.
* **Suspicion of Mass Mailing**: This email may be a mass mailing (newsletter, promotion, or commercial communication) you did not opt into. Do not click any link, including an unsubscribe link, if you do not recognize the sender. If in doubt, ignore this message and do not share any information.
* **Suspicion of Deceptive Commercial Spam**: This email may be an aggressive commercial message using deceptive tactics (fake urgency, misleading offer, spoofed sender) to get you to click, buy, or share information. Do not click any link, do not open any attachment, and do not share any data. If in doubt, ignore this message and report it to your IT manager.
* **Suspicion of Spam** (fallback, used when no specific type of spam is detected): This email may be an unsolicited commercial message. Do not reply, do not click any link, and do not open any attachment if you were not expecting this message. If in doubt, ignore it and report it to your IT manager.

</details>

<details>

<summary>Generic</summary>

* Used when no specific category (fraud, phishing, malware, spam) is detected: This email has been flagged as suspicious. As a precaution, do not click any link, do not open any attachment, and do not share any information. If in doubt, verify with the sender through a trusted channel before taking any action.

</details>

### Quarantine

The email is moved to a dedicated `Stoïk Quarantine` folder. The employee no longer sees it in their inbox, and it remains recoverable.

## Compromised accounts

When an account is identified as compromised, the Stoïk SOC can:

* **Reset password**: issues a temporary password that must be changed at the next login.
* **Reset session**: revokes all active sessions, forcing re-authentication on every device.
* **Delete inbox rules**: removes a mailbox rule (forwarding, auto-deletion, etc.).
* **Delete applications**: revokes an application or OAuth grant.
* **Delete MFA method**: removes an authentication method.
* **Delete device**: permanently deletes a device registered on the account.
* **Delete passkey**: removes a passkey added by an attacker.


# Link analysis

Email Security inspects the links found in the emails it analyzes. Here is what our systems do with those links, and how to report a link that should be left alone.

## Why links are resolved

Phishing rarely gives itself away in the body of a message. The signal is in the destination: a shortened link that lands on a lookalike domain, or a redirect chain that ends on a domain registered three days ago.

Reading a link as it is written tells us none of that. The link has to be resolved to see where it actually goes.

## What the analysis does

For each link in an analyzed email, Stoïk sends one HTTP `HEAD` request and follows up to five redirects.

A `HEAD` request asks a server for the response headers only. It is the same request as opening the page, minus the page content. It is used instead of a full page load because it is the lightest way to learn where a link leads.

From it, Stoïk records the end of the redirect chain, the number of hops, and the reputation of the destination domain. The request times out after two seconds.

## When the analysis goes further

If the destination domain looks suspicious, meaning it has a poor reputation, was registered recently, or is rarely seen across the mailboxes we protect, the page is opened in an isolated browser environment and a screenshot is captured. A vision model then judges whether the page is a credential harvesting kit, and which brand it imitates.

A domain that is well established and commonly seen never reaches that stage. A link to a major provider is resolved, but its page is never loaded.

Loading a page means the page runs its own JavaScript, exactly as it would in a visitor's browser, and a page can issue requests of its own while it does. Those requests come from the page, not from instructions we give it. When a page is unreachable because of an anti-bot challenge, the capture goes through a solver that negotiates the challenge first.

## What the analysis never does

Stoïk does not act on your behalf. Our analysis clicks no button, fills in no form, and never uses your credentials. Where a link would trigger something by the mere fact of being requested, it is not requested at all: see [Links that are skipped entirely](#links-that-are-skipped-entirely).

### Exception: file-sharing platforms

Attackers run a growing share of their phishing through legitimate file-sharing services. A document shared from a well-known platform borrows that platform's reputation and gets past most filters, while the trap sits inside the shared file. Judging that email means reading what was actually shared.

These platforms do not hand a shared document to a plain page load. Reaching one means making the request the platform defines for it: a query to its own API, or the anonymous authentication step a share link performs using the key already contained in the link.

Stoïk does this only on the platforms it explicitly supports, and only with what the link already carries.

## Links that are skipped entirely

Some links change something the moment they are requested, even by a `HEAD`. An electronic signature link can mark a document as viewed or consume a one-time token. A quarantine digest link can release a blocked email. A one-click unsubscribe link takes the recipient off a mailing list, and a calendar invitation link can record an RSVP.

The HTTP standard is explicit here. RFC 9110 section 9.2.1 defines `GET` and `HEAD` as safe methods: a client sending one neither requests nor expects any change of state on the server. The same section adds that when a parameter in the URL selects an action, the owner of that resource must disable the action when it is reached with a safe method. The rule exists so that automated retrieval can work without causing harm.

Not every service follows it.

{% hint style="info" %}
Stoïk keeps an allowlist of destinations that are skipped completely, with no request sent at all. It covers electronic signature platforms, email quarantine portals, newsletter unsubscribe endpoints, and calendar invitation responses.
{% endhint %}

## Reporting a link

If you see an action on one of your services that you believe came from our analysis, report it to Stoïk through your usual support channel.

Useful details:

* The service concerned
* The address of the mailbox affected
* Roughly when it happened
* The link itself, if you still have it

The destination is added to the allowlist, and you are notified once the change is live.


# Blocklist

Block a sender or a whole domain yourself, without contacting us. Use it when a sender keeps sending unwanted emails, or to stop all emails from a domain, whatever the sender.

## Add a rule

1. In Stoïk Protect, go to `Suspicious emails` > `Settings` and open the `Blocklist` tab.
2. Click `+ Add`.
3. Choose what to block: a `Sender` or a `Domain`, and enter it.
4. Pick the tenants where the rule applies.
5. Optionally, add a note to remember why you blocked it.
6. Save the rule.

The rule applies only to the tenants you selected.

## What happens to a blocked email

* It goes to quarantine, with a generic suspicious banner added to it.
* No alert is raised in Stoïk Protect.

## Edit or delete a rule

Edit or delete a rule at any time from the `Blocklist` tab. Each rule shows the tenants it applies to and, if you added one, the note explaining why it exists.

If an email you expected to be blocked still reaches your inbox, write to us at <protect@stoik.io>.




---

[Next Page](https://docs.stoik.io/llms-full.txt/1)

