Topic Hub

Endpoint Security

Hardening baselines, Defender for Endpoint, BitLocker, LAPS, and Conditional Access enforcement. Practical guidance for securing Windows endpoints in Microsoft-managed environments — from first boot through ongoing operations.

Guides, scripts and analysis

Overview

What Endpoint Security Covers in Windows Environments

Endpoint security is not a single product — it is a stack of overlapping controls. In a Microsoft environment, that stack runs from hardware-level Secure Boot through identity-linked Conditional Access, with Intune and Defender providing the management and detection layers in between.

Hardening baselines

Applying a defined security configuration — CIS Benchmark, Microsoft Security Baseline, or DISA STIG — reduces the attack surface by disabling unused services, enforcing audit policy, restricting credential exposure, and enabling exploit mitigations.

Antivirus and EDR

Microsoft Defender Antivirus is the built-in AV layer. Defender for Endpoint (MDE) adds behavioural detection, threat intelligence, automated investigation, and the advanced hunting query interface. Third-party EDR (CrowdStrike, SentinelOne) replaces or supplements these capabilities.

Device encryption

BitLocker encrypts the OS volume using TPM-backed key protection. Proper deployment requires Secure Boot, a TPM 2.0 chip, and key escrow to Entra ID or Active Directory before encryption is enforced — so keys are retrievable when users are locked out.

Local administrator control

Unmanaged local admin accounts are a persistent lateral movement risk. Windows LAPS rotates the local administrator password automatically and stores it in Entra ID or AD, making the credential unique per device and auditable.

Conditional Access and compliance

Entra Conditional Access uses device compliance state from Intune as a gate for cloud resource access. Non-compliant or unmanaged devices can be blocked from Microsoft 365, SharePoint, and other services — without VPN.

Vulnerability and patch exposure

Unpatched endpoints are the most common initial access vector. Defender Vulnerability Management scores exposure per device based on installed software, OS version, and configuration weaknesses — and integrates with Intune for remediation tracking.

Hardening

Windows Hardening Baseline

A hardening baseline is a documented, repeatable configuration applied to every managed endpoint. Three major baselines are in common enterprise use — each with different scope, update cadence, and compliance mapping requirements.

Most widely adopted

CIS Benchmark — Level 1

What it covers
Account policies, local policies, Windows Firewall, audit policy, user rights, and security options. Level 1 is intended for broad deployment without impacting usability.
Level 2 additions
More aggressive restrictions — AppLocker, Windows Installer restrictions, SMBv1 disablement, and stricter audit policy. May require application compatibility testing before broad rollout.
Deployment method
GPO import (CIS provides a GPO pack), Intune Settings Catalog, or a PowerShell script such as Invoke-WindowsHardening running as a Proactive Remediation.
Validation
CIS-CAT Pro or Invoke-WindowsHardening with reporting mode generates a per-control pass/fail delta against the current state.

Built for M365 environments

Microsoft Security Baseline

What it covers
Microsoft's recommended settings for Windows 11 and Windows Server — aligned to their own Secure Score framework. Delivered as GPO templates and Intune configuration profile exports.
Tooling
Download via the Microsoft Security Compliance Toolkit. Includes Policy Analyzer for diffing baseline against current GPO state and LGPO.exe for local policy import.
Update cadence
Released per Windows feature update cycle. The baseline for 24H2 differs from 23H2 — review the changelog before applying to a new OS ring.
Intune integration
Settings Catalog templates in Intune map directly to many baseline controls. The Endpoint security > Security baselines node provides a pre-built Intune profile based on the Microsoft baseline.

Regulated / government

DISA STIG

What it covers
The most prescriptive of the three. Covers every configurable security setting with a Finding ID, severity category (CAT I/II/III), and fix text. Used in US federal and regulated sectors.
Application scope
Separate STIGs for Windows 11, Windows Server 2022, Defender AV, Edge, and Office. Each must be applied and assessed independently.
Tooling
STIG Viewer (DISA) for reading findings. OpenSCAP or SCC (SCAP Compliance Checker) for automated assessment. GPO packages are available from DISA for each STIG.
Trade-off
Full STIG compliance often conflicts with usability in standard enterprise environments. Tailor findings with an accepted risk documentation process rather than applying everything wholesale.

Attack Surface Reduction rules

ASR rules block specific attack techniques at the OS level — Office macro spawning child processes, credential theft from LSASS, and executable content from email. Audit mode before enforcement to find false positives.

Credential Guard

Isolates LSASS credentials in a Hyper-V protected container, preventing pass-the-hash and credential dump attacks. Enabled via Intune device configuration or GPO. Requires UEFI with Secure Boot and 64-bit Windows.

SMB and legacy protocol hardening

Disable SMBv1 (disabled by default in Windows 11 but verify on older builds), enforce SMB signing, disable NetBIOS over TCP/IP, and restrict NTLMv1 via security policy — all standard CIS Level 1 controls.

LSASS protection

Enable RunAsPPL (Protected Process Light) for LSASS via registry or Intune CSP. Prevents credential dumping tools from accessing LSASS memory. Verify driver compatibility before fleet-wide rollout.

Detection and response

Defender, EDR, and Third-Party Security Tooling

Microsoft Defender for Endpoint is the primary EDR platform for Microsoft-managed environments. How you deploy and configure it — and whether you supplement it with a third-party tool — depends on your licensing, threat model, and operational maturity.

Microsoft Defender for Endpoint (MDE)

Behavioural detection engine, threat intelligence, automated investigation and response (AIR), and the advanced hunting interface. Available in Plan 1 (device-based controls, ASR, firewall management) and Plan 2 (full EDR, threat analytics, vulnerability management). Included in Microsoft 365 E5 Security.

Deploying MDE via Intune

Onboard devices via Intune using the Endpoint security > Endpoint detection and response policy. The onboarding package is generated in the Defender portal and deployed as a configuration profile. Verify onboarding status in the Defender portal under Devices — a device should appear within 24 hours of policy assignment.

Defender Antivirus exclusions

Over-exclusion is one of the most common ways MDE is weakened in production. Exclusions should be path-specific and time-limited where possible, documented, and reviewed quarterly. Broad folder exclusions (e.g., C:\) are never appropriate outside of isolated test environments.

Attack Surface Reduction in production

Run ASR rules in Audit mode for at least two weeks before switching to Block. Review the ASR audit event log (Event ID 1121/1122 in Microsoft-Windows-Windows Defender/Operational) and the MDE advanced hunting table DeviceEvents to identify false positives before enforcement.

Third-party EDR (CrowdStrike, SentinelOne)

When a third-party EDR is deployed, Defender AV runs in passive mode — it scans but does not block independently. The third-party sensor owns the active prevention role. Confirm passive mode is set correctly via the registry key ForceDefenderPassiveMode or Intune policy to avoid dual-blocking conflicts.

Defender Vulnerability Management

Available in MDE Plan 2, DVM scores each device's exposure based on installed software CVEs, OS vulnerabilities, and configuration weaknesses. Security recommendations link directly to Intune remediation tasks. The Exposure Score aggregates fleet-wide risk into a single trackable metric.

Encryption

BitLocker and Device Encryption

BitLocker protects data at rest on lost or stolen devices. The silent failure mode — encryption enabled but keys not escrowed — is the most consequential misconfiguration. Design your policy around escrow verification, not just encryption state.

TPM and hardware requirements

BitLocker with TPM binding provides the strongest protection — the key is sealed to the TPM and released only when boot measurements match the expected PCR values. Requires TPM 1.2 minimum; TPM 2.0 is required for Autopilot and Windows 11 compliance policies. Devices without a TPM can use a USB startup key, but this is not suitable for unattended deployment.

Intune BitLocker policy

Configure via Endpoint security > Disk encryption or via a Settings Catalog profile. Key settings: require encryption on OS drive, enforce TPM-only or TPM+PIN protector, configure recovery key backup to Entra ID, and set the encryption method (XTS-AES 128 or 256). Require device encryption in the compliance policy to gate Conditional Access.

Key escrow to Entra ID

Keys escrow silently when the BitLocker policy applies and the device is Entra joined or hybrid joined. Escrow failure is silent — no error appears on the device or in Intune. Verify escrow by checking Entra ID > Devices > [device] > Recovery Keys. Force re-escrow using Manage-bde -protectors -adbackup C: if keys are missing.

PCR7 binding and patch risk

Patches that modify Secure Boot databases or UEFI CA certificates can break PCR7 binding, triggering BitLocker recovery on next boot — even when encryption and keys are configured correctly. Suspend BitLocker for one reboot before applying high-risk patches (especially those involving Secure Boot or UEFI) to pilot rings first.

Recovery workflow

When a user hits a recovery screen: retrieve the 48-digit recovery key from Entra ID (Devices > [device] > Recovery Keys) or from on-prem AD (using the BitLocker Recovery Password Viewer in ADUC). After recovery, investigate why recovery was triggered — PCR7 change, TPM clear, or policy conflict — before the device returns to service.

Device Encryption vs BitLocker

Consumer editions of Windows use Device Encryption (a simplified BitLocker subset) which activates automatically on compatible hardware. Enterprise environments should use full BitLocker with explicit policy to control protector type, algorithm, and key escrow. Device Encryption does not appear in the Intune Disk encryption report.

Local admin control

Windows LAPS and Local Administrator Management

An unmanaged local administrator account with a shared or default password is one of the highest-impact lateral movement risks in an enterprise network. Windows LAPS (Local Administrator Password Solution) removes shared credentials by making every local admin password unique, rotated, and auditable.

What Windows LAPS does

LAPS rotates the local administrator password on a configurable schedule (default: 30 days) and stores the encrypted password in Entra ID or on-premises Active Directory. Admins retrieve passwords per-device via the Intune portal, Entra ID device blade, or the Get-LapsAADPassword PowerShell cmdlet — with retrieval logged for audit.

Deploying LAPS with Intune

Enable via Endpoint security > Account protection > Local admin password solution policy. Configure the managed account name, password complexity, password length (minimum 14 characters recommended), and rotation period. Verify deployment status per-device in the Intune device configuration blade.

Native LAPS for Entra-joined devices

Windows LAPS (built into Windows 11 22H2+ and Windows Server 2025) supports Entra ID as the backup directory natively — no on-premises AD required. Keys are backed up using the Entra ID device object. This replaces the previous requirement for hybrid join or the legacy Microsoft LAPS client.

Migrating from legacy Microsoft LAPS

Legacy LAPS (the separate download) stores passwords in a plaintext AD attribute and requires domain join. If you have legacy LAPS deployed on hybrid-joined devices, migrate to Windows LAPS by deploying the new Intune LAPS policy, which will take over management of the designated account. Remove the legacy LAPS GPO once the new policy is confirmed active.

Account naming and scope

Target the built-in Administrator account (SID S-1-5-21-...-500) or a named local account. Do not target all local accounts — LAPS manages one account per policy. If your image creates a custom admin account, specify that account name in the policy configuration.

Audit and access control

Every password retrieval from Entra ID generates an audit log entry. Configure Entra ID diagnostic settings to export these to a Log Analytics workspace for retention beyond the default 30-day portal window. Restrict password retrieval to specific admin roles — avoid granting Global Admin or Intune Admin roles solely for LAPS retrieval.

Access control

Conditional Access and Device Compliance

Compliance policy and Conditional Access form the enforcement layer — they translate device security state into access decisions for cloud resources. A device that does not meet the compliance baseline loses access to Microsoft 365 without VPN or firewall changes required.

Intune compliance settings for security

Key compliance checks for a security baseline: minimum OS version, BitLocker enabled, Secure Boot on, code integrity on, Defender real-time protection active, and maximum allowed threat severity from Defender for Endpoint set to Medium or lower.

Conditional Access: Require compliant device

The "Require device to be marked as compliant" grant control in Entra Conditional Access blocks access from any device not registered in Intune or not meeting the compliance baseline. Apply to Microsoft 365 apps at minimum. Exclude emergency access (break-glass) accounts from all CA policies.

Noncompliance grace period

Set a grace period (typically 1–3 days) before a device is officially marked noncompliant. This gives users and IT time to respond to a compliance drift before access is blocked. For high-security environments, set grace period to 0 for immediate blocking.

Named locations and network-based CA

Trusted IP ranges (office networks, VPN exit nodes) can be defined as Named Locations in Entra. CA policies can require compliant device OR trusted location — reducing friction for desk-based users while enforcing MDM enrollment for remote access.

Sign-in risk and user risk policies

Entra ID Protection (P2 licence) generates risk signals based on impossible travel, leaked credentials, and anomalous sign-in patterns. Risk-based CA policies require MFA re-authentication or password change when elevated risk is detected — without requiring admin intervention.

Monitoring compliance drift

Intune Devices > Monitor > Noncompliant devices shows current noncompliance by reason. The Conditional Access Insights and Reporting workbook in Entra shows which policies are blocking sign-ins and for which users — essential for validating policy before enabling enforced mode.

Vulnerability exposure

Patching and Vulnerability Exposure

Unpatched endpoints are the most frequently exploited initial access vector. Patch compliance reporting and vulnerability management close the gap between what is deployed and what is exposed.

CVSS vs EPSS for prioritisation

CVSS (Common Vulnerability Scoring System) rates severity — but high-CVSS vulnerabilities are not always actively exploited. EPSS (Exploit Prediction Scoring System) estimates the probability a CVE will be exploited in the wild within 30 days. Combining both gives a better triage signal than CVSS alone.

Defender Vulnerability Management

MDE Plan 2 includes Defender Vulnerability Management, which scores each device's software inventory against the CVE database and flags configuration weaknesses. Security recommendations appear in the Defender portal and can be pushed as Intune remediation tasks.

Patch ring timing relative to CVEs

CVSS ≥ 9.0 or CISA KEV listed? Deploy to all rings within 14 days of patch availability — compress the standard ring schedule. Actively exploited zero-days may warrant emergency out-of-band deployment to production if compensating controls are insufficient.

Intune update compliance

Intune Devices > Monitor > Feature update failures and Quality update compliance show per-device patch state. Devices that have not applied a quality update within the deadline window appear as non-compliant — combine with the compliance policy to gate CA access.

End-of-support exposure

Devices running Windows versions past end-of-support (Windows 10 21H2 reached EOS November 2023) receive no security patches. Defender Vulnerability Management flags EOS OS versions as high-risk. Prioritise feature update rollout for devices on EOS builds.

Software inventory gaps

LOB applications installed outside of Intune app management may not appear in the Defender software inventory. Use Export-IntuneDeviceReport combined with Get-StaleDevices to cross-reference managed vs detected software, identifying unmanaged installs that may carry CVE exposure.

Monitoring

Monitoring, Reporting, and Incident Response Workflow

Detection is only useful if you act on it quickly. These are the primary data sources, tools, and workflows used to detect, investigate, and contain endpoint security incidents in Microsoft environments.

Defender portal — Incidents queue

security.microsoft.com > Incidents & alerts > Incidents. MDE correlates individual alerts into incidents based on shared entities (device, user, file hash). Start investigation at the incident level, not the alert level — the incident graph shows the full attack chain and reduces alert noise.

Advanced Hunting (KQL)

The Advanced Hunting interface in the Defender portal accepts KQL queries across DeviceProcessEvents, DeviceNetworkEvents, DeviceFileEvents, DeviceLogonEvents, and other tables. Queries can be scheduled as Custom Detection rules that generate alerts when matching events occur.

Windows Event log — Security channel

Event ID 4624 (logon success), 4625 (logon failure), 4648 (logon with explicit credentials), 4688 (process creation with command line logging enabled), and 4698/4702 (scheduled task created/modified). Enable command-line auditing via GPO or Intune for 4688 to be useful in incident response.

Microsoft Sentinel integration

MDE data can be forwarded to Microsoft Sentinel via the Microsoft Defender XDR connector. Sentinel provides extended retention (up to 12 years), cross-product correlation (MDE + Entra + Defender for Cloud), and SOAR playbooks for automated response actions.

Live Response in MDE

MDE Plan 2 Live Response provides a remote shell into managed devices without requiring RDP or a VPN tunnel. Use it to collect forensic artefacts, run investigation scripts, or isolate a device from the network. All Live Response sessions are logged and auditable in the Defender portal.

Device isolation and containment

MDE can isolate a compromised device from the network (blocking all traffic except to the Defender cloud and the configured management channel) from the portal or via API. The device remains manageable through Intune and Defender during isolation. Confirm with the user before isolating shared or kiosk devices.

Common problems

Where Endpoint Security Deployments Go Wrong

Most endpoint security failures are configuration gaps rather than technology limits. These are the patterns that appear most often in production environments.

BitLocker enabled but keys not escrowed

Encryption shows as active in the Intune compliance blade but the recovery key is absent from Entra ID. This is a silent failure — no alert is raised. Verify key escrow by checking Entra ID device properties for each enrolled device. Force re-escrow with Manage-bde -protectors -adbackup C: on affected devices.

ASR rules causing application breakage

Attack Surface Reduction rules in Block mode can silently kill LOB application functionality — Office macro spawning processes, custom update mechanisms, or signed scripts that match a rule signature. Always run Audit mode for two weeks and review DeviceEvents in advanced hunting before switching to Block.

Defender in passive mode without a replacement AV

When a third-party AV is installed, Defender switches to passive mode and stops actively blocking. If the third-party AV is later uninstalled or disabled, Defender may not automatically return to active mode — leaving the device with no active AV protection. Verify Defender state after any AV software changes.

Compliance policy missing a device — default is compliant

Intune's default compliance stance for devices with no assigned policy is "Compliant." A device that falls through group targeting gaps gets a free pass to Conditional Access-gated resources. Audit group membership regularly and set the default to "Not Compliant" in Endpoint security > Compliance policy settings if appropriate for your environment.

LAPS policy deployed but wrong account targeted

LAPS manages one account per policy. If the policy targets "Administrator" but your image renames the built-in account or creates a different admin account, the policy applies to a non-existent account and no rotation occurs. Confirm the target account name matches what actually exists on the device using Get-LocalUser.

CIS baseline applied without application compatibility testing

CIS Level 2 controls — AppLocker, strict NTLM restrictions, or UAC settings — can break LOB application workflows that developers or vendors never tested against a hardened baseline. Test Level 1 first across your application profile before adding Level 2 controls, and document any exceptions.

MDE not fully onboarded — device appears in Intune but not Defender

The Intune MDM enrollment and the MDE onboarding are separate processes. A device enrolled in Intune will not appear in the Defender portal until the onboarding policy is applied and processed. Check Endpoint security > Endpoint detection and response policy assignment status for devices not appearing in Defender.

Exclusion scope too broad

Broad Defender AV exclusions — entire directory trees, all processes from an application vendor, or specific file extensions applied globally — create blind spots exploited by attackers who know the exclusion patterns. Scope exclusions to the minimum path or process required, document the business justification, and review quarterly.