Ransomware that encrypts data, wipers that destroy it and insiders who quietly alter it all attack the same property: whether the organisation can trust its own data. NIST's National Cybersecurity Center of Excellence (NCCoE) built three practice guides on data integrity: SP 1800-25 (identify and protect), SP 1800-26 (detect and respond) and SP 1800-11 (recover). Together they describe the capabilities, example architectures and test scenarios for keeping business data, systems and configurations intact and restorable. This guide explains what the trilogy contains, how its capabilities fit together, where identity and machine credentials matter, and how to turn it into an implementation and assessment plan. It is written for security architects, infrastructure leads and resilience teams. It complements the Ransomware Response and Recovery Guide, which covers the incident itself.
Key takeaways
- Three guides, one lifecycle: SP 1800-25 identifies and protects assets, SP 1800-26 detects and responds to integrity events, SP 1800-11 recovers and confirms the recovered data can be trusted.
- Integrity monitoring is the thread through all three: baselines of files, systems, databases and directory objects let you see what changed, when, and who or what changed it.
- Backups only count if attackers cannot alter them: the guides pair backup capability with secure, write-once storage for backups, logs and golden images.
- Several test scenarios start with an identity: backdoor accounts, stolen administrator credentials and a maintenance script deleting a virtual machine, so account and service identity controls are part of data integrity.
- These are practice guides, not mandates: they show example builds with named products and map to CSF 1.1 subcategories, so you adapt the capabilities rather than copy the bill of materials.
What it is and who it applies to
The NIST SP 1800 series are NCCoE practice guides: each documents a problem, an example solution built in the NCCoE lab with technology collaborators, and how-to material. They are voluntary. Organisations use them to design architectures, justify investment and test their own capability against realistic scenarios. Each guide has three volumes: A (executive summary), B (approach, architecture and security characteristics) and C (how-to guides).
| Publication | Full title | Status and date (checked 27 Oct 2026) | Focus |
|---|---|---|---|
| SP 1800-25 | Data Integrity: Identifying and Protecting Assets Against Ransomware and Other Destructive Events | Final, December 2020 | Know which devices, data and applications exist, find their vulnerabilities, protect them from corruption and destruction |
| SP 1800-26 | Data Integrity: Detecting and Responding to Ransomware and Other Destructive Events | Final, December 2020 | Detect, mitigate and contain data integrity events quickly to reduce downtime and loss |
| SP 1800-11 | Data Integrity: Recovering from Ransomware and Other Destructive Events | Final, September 2020 | Recover operating systems, databases, user files, applications and configurations, and trust the result |
The threats in scope are ransomware, destructive malware, malicious insiders and honest user mistakes. That last item is often overlooked: an accidental deletion by a script or an AI coding agent is an integrity event too, as the Replit AI agent database deletion showed, where rollback restored the data even after the agent claimed it could not. The guides apply to any enterprise IT environment; the related NCCoE data confidentiality projects, including SP 1800-28, cover data breaches rather than integrity.
Structure and key requirements
Capabilities across the trilogy
| Capability | Guide | What it does in the reference design |
|---|---|---|
| Inventory | 1800-25 | Discovers and tracks users, devices, software and data assets, including sensitive data |
| Vulnerability management | 1800-25 | Scans hosts for known vulnerabilities and prioritises by risk |
| Policy enforcement | 1800-25 | Posture-checks machines and blocks network access for non-compliant ones |
| Integrity monitoring | 1800-25, 1800-26 (and corruption testing in 1800-11) | Baselines files, systems and Active Directory; detects modifications, deletions and additions and who made them; supports granular directory rollback |
| Logging | All three | Collects and correlates logs from Windows, Linux, file integrity, database queries and other sources; establishes activity baselines and supports audit |
| Backups | 1800-25, 1800-11 | Encrypted, scheduled backups of data, directory, systems and configurations; file, disk, system and whole-VM restore |
| Secure storage | 1800-25, 1800-11 | Write-once (WORM) storage for backups, logs, configurations and golden images that cannot be altered once locked |
| Network protection and denylisting | 1800-25 | Restricts communication to limit lateral movement; blocks known-malicious sites and addresses through a proxy |
| Event detection | 1800-26 | Detects malicious software, malicious email attachments, suspicious network activity and directory changes as they happen |
| Mitigation and containment | 1800-26 | Removes malicious files, quarantines devices, reverts unauthorised directory changes, disables accounts, sanitises attachments |
| Forensics and analytics | 1800-26 | Correlation and traffic analysis to find source and scope |
| Reporting | 1800-26 | Alerts and reports on enterprise health and malware detection |
| Virtual infrastructure | 1800-11 | Backs up and restores whole virtual machines and individual guest files |
SP 1800-11 also protects the solution itself with a management network and firewall-isolated subnets. That design choice is easy to skip and important: if the backup and logging systems share credentials and networks with production, one compromise reaches both.
Test scenarios
The guides test their builds against concrete scenarios, which make useful exercise scripts. SP 1800-25 and 1800-26 share seven: ransomware via web download that spreads through privilege escalation; destructive malware on a USB drive; accidental VM deletion by a maintenance script; a backdoor account created via a malicious email attachment; database modification by an insider through a web vulnerability; file and backup modification by an insider using stolen administrator credentials and remote PowerShell; and a backdoor delivered by a compromised update server. SP 1800-25 adds a new employee scenario for inventory and policy enforcement. SP 1800-11 covers ransomware encryption, file modification and deletion, VM deletion, Active Directory permission change with backdoor accounts, harmful database transactions and database metadata modification.
How the guides fit together
Read them as one control loop rather than three documents. SP 1800-25 establishes what "normal" looks like, through inventory and baselines, and reduces the chance of an event. SP 1800-26 uses those same baselines to notice deviation quickly and to act on it: the integrity monitoring tools appear in both builds for that reason. SP 1800-11 then depends on both: a restore is only trustworthy if you know the last good state and can prove the restored copy matches it. A gap in one stage weakens the others. Without baselines, detection produces noise and recovery has nothing to verify against; without protected backups, fast detection only tells you sooner that data is gone. When planning budget, fund baselines, protected storage and logging first, because every later capability relies on them.
Identity and NHI requirements
The trilogy is not an identity standard, but its scenarios repeatedly turn on accounts. Map them to identity controls explicitly:
- Backdoor accounts: two scenarios create unapproved accounts (T1136 Create Account). Integrity monitoring of the directory, plus the ability to revert unauthorised directory changes and disable accounts, is the documented response. See the Identity Recovery Guide.
- Stolen administrator credentials: the insider scenario alters both data and backups using administrator credentials and remote PowerShell (T1078 Valid Accounts). Backup administration must use separate identities from production administration. See the Backup Identity Isolation Guide.
- Automation and scripts: the VM deletion scenario is a maintenance script, a non-human identity acting with too much authority. Scope automation credentials to the objects they maintain and log their actions separately.
- Management tooling: device management, software deployment and update servers can deliver destruction at scale (T1072 Software Deployment Tools). In the Stryker Intune wiper attack, attackers compromised an administrator account, created a new Global Administrator and issued remote wipe commands across the device fleet.
- Service accounts for backup, logging and integrity agents: these hold broad read or write rights by design. Inventory them, vault their credentials and alert on interactive use. See the Service Account Security Guide.
Implementation steps
- Define integrity-critical assets. Use the business impact analysis to list systems and data whose corruption would stop the business: directory, ERP and finance databases, file shares, configuration repositories, virtual infrastructure. Record owners.
- Inventory and baseline (1800-25). Build the device, software and data inventory; capture known-good baselines of files, configurations, database metadata and directory objects; store golden images.
- Close the obvious paths. Vulnerability management with risk-based prioritisation, network segmentation for management planes, and denylisting at the web proxy.
- Protect backups. Encrypted, scheduled backups of data, directory, systems and configuration; a copy on immutable or write-once storage; backup infrastructure on an isolated management network with separate identities. See the Backup Security and Immutable Backup Guide.
- Detect (1800-26). Turn integrity monitoring changes, directory changes, endpoint malware detections and network anomalies into correlated SIEM use cases with owners and response steps.
- Pre-authorise containment. Decide in advance who may quarantine devices, disable accounts and revert directory changes, and test those actions.
- Recover and verify (1800-11). Document restore procedures per asset type: files, databases (including transaction rollback), whole VMs, directory system state. Compare restored data against integrity baselines before returning it to service. See the Restore Testing and Recovery Assurance Guide.
- Exercise the scenarios. Run the NCCoE scenarios as tabletop and technical tests, starting with VM deletion and file modification, which are low risk to simulate in a lab.
Rollback and safety: integrity monitoring and automated containment can disrupt operations if tuned poorly. Start in alert-only mode, measure false positives for at least one change cycle, then enable automated actions for the highest-confidence detections only.
Choosing an approach
| Organisation | Practical starting point |
|---|---|
| Small, mostly SaaS | Rely on provider versioning and retention, add a third-party backup of SaaS data with separate admin identities, and monitor admin role changes |
| Mid-size hybrid | Immutable backup copy, directory change monitoring and rollback, file integrity monitoring on servers holding critical data, SIEM correlation |
| Large or regulated | Full trilogy capability set, isolated recovery environment, integrity verification before restore, annual scenario-based testing with evidence |
Evidence and assessment
Because the guides are not certifiable, assessment means showing that each capability exists and works. Useful evidence:
- Inventory: current asset and data inventory with owners and integrity-critical flags.
- Baselines: dated baselines for critical servers, databases and directory, and change reports reconciled with change tickets.
- Backups: job success reports, immutability settings, and the identities that can change retention.
- Detection: SIEM rules for mass file modification, backup deletion, directory privilege changes and new admin accounts, with test results.
- Recovery: restore test records per asset type, with measured recovery time and integrity check results.
Metrics that show the programme working:
- Percentage of integrity-critical assets with a current baseline and monitoring.
- Percentage of critical data with an immutable backup copy.
- Restore success rate and measured recovery time against target, per asset type.
- Mean time to detect unauthorised directory or backup changes in exercises.
- Number of accounts able to delete or change backup retention, trended down.
Mapping to other frameworks
The NCCoE guides map their capabilities to NIST CSF 1.1 subcategories (for example PR.DS-6 integrity checking and PR.IP-4 backups). CSF 2.0 reorganised these, so use the 2.0 identifiers below for new work. Mappings in the table are the authors' alignment, not NIST's.
| Capability | NIST CSF 2.0 | NIST SP 800-53 Rev. 5 | MITRE ATT&CK threats addressed |
|---|---|---|---|
| Inventory | ID.AM-01, ID.AM-02, ID.AM-07 | CM-8 | Supports all |
| Vulnerability management | ID.RA-01 | RA-5 | Initial access and spread |
| Integrity monitoring and corruption testing | DE.CM-09 | SI-7 | T1565.001 Stored Data Manipulation |
| Logging and analytics | PR.PS-04, DE.AE-02, RS.AN-03 | AU-2, AU-6, SI-4 | All |
| Backups and secure storage | PR.DS-11 | CP-9 | T1486, T1485, T1490 |
| Network protection | PR.IR-01 | SC-7 | Lateral movement |
| Mitigation and containment | RS.MI-01 | IR-4 | T1561 Disk Wipe, T1486 |
| Recovery and verification | RC.RP-03, RC.RP-05 | CP-10 | All impact techniques |
For a ransomware-specific CSF profile, see the NIST IR 8374 Rev 1 Ransomware Risk Management Profile Guide. For wipers specifically, see the Wiper Malware Guide.
Common pitfalls
- Copying the product list: the example builds name specific products and versions from around 2020. Adopt the capabilities and test scenarios; select current tools on your own criteria.
- Backups in the same trust domain: backup servers joined to the production directory and managed with the same admin accounts can be deleted by the same attacker, which is exactly the insider scenario the guides test.
- Restoring without verifying: restoring corrupted or backdoored data quickly makes the incident longer. SP 1800-11's emphasis is on recovered data that can be trusted.
- Ignoring the directory: several scenarios alter Active Directory. Without directory change monitoring and rollback, file and VM restores leave backdoor accounts in place.
- Logs that the attacker can delete: logs belong on protected storage, which is why the guides list logs alongside backups for secure storage.
- Treating accidents as out of scope: script and agent mistakes cause real integrity loss and are cheaper to exercise than ransomware.
Practitioner checklist
- List integrity-critical systems and data with owners, using the business impact analysis.
- Capture baselines for critical files, configurations, database metadata and directory objects.
- Keep at least one immutable or write-once copy of backups, logs and golden images.
- Run backup and logging infrastructure on isolated networks with separate administrator identities.
- Monitor and alert on directory privilege changes, new admin accounts and backup retention changes.
- Pre-authorise containment actions and test reverting unauthorised directory changes.
- Verify restored data against baselines before returning systems to service.
- Exercise the NCCoE scenarios at least annually and record recovery times.
- Scope automation and AI agent credentials so they cannot delete production data unaided.
Standards and references
- NIST: SP 1800-25, Data Integrity: Identifying and Protecting Assets Against Ransomware and Other Destructive Events
- NIST: SP 1800-26, Data Integrity: Detecting and Responding to Ransomware and Other Destructive Events
- NIST: SP 1800-11, Data Integrity: Recovering from Ransomware and Other Destructive Events
- NIST NCCoE: Data Integrity project page
- NIST: Cybersecurity Framework 2.0
- NIST: SP 800-53 Rev. 5
- MITRE ATT&CK: Data Destruction
- MITRE ATT&CK: Stored Data Manipulation
Related NHI Mgmt Group resources: Ransomware Response and Recovery Guide · Wiper Attack Response Playbook · Cybersecurity Event Recovery Guide · Cyber Recovery Vault Guide