Join our Newsletter — 33% off our NHI Course
Home› Guides› D3-RS Restore Software: Reinstalling Trusted Software and Verifying…
Guide Resilience & Recovery

D3-RS Restore Software: Reinstalling Trusted Software and Verifying the Restore

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 12 November 2026 11 min read
On this page

Software restored from the wrong source reinstalls the compromise you just removed. D3-RS Restore Software is the MITRE D3FEND countermeasure for putting software back on a host after it was tampered with, removed or replaced during an incident. D3FEND defines it as "Restoring software to a host", places it in the Restore tactic under Restore Object (D3-RO) and cites CISA's Cybersecurity Incident and Vulnerability Response Playbooks as its reference. D3FEND gives no implementation article, so this card covers the practical questions: which source to trust, the repair and reinstall mechanisms operating systems provide, how to verify the restored software, and the limits. It is written for incident responders, endpoint and server engineers and the teams that run software distribution.

Key takeaways

  • D3-RS returns software to a trusted state. It covers reinstalling or repairing applications, operating system components, add-ins, server modules and agents after tampering or removal.
  • The source decides the outcome. Reinstall from signed vendor packages, a verified internal repository or a gold image, never from a copy on the compromised host.
  • Supply chain cases need a different restore point. If the vendor package itself was compromised (T1195.002), restore a version the vendor has confirmed clean, not the latest one.
  • D3FEND's ATT&CK links are inferred from the Software artefact. They point to where tampered software hides: add-ins, IIS modules, SQL stored procedures, installers, firmware and deployment tools.
  • Verify with the platform's own integrity tools. Package manager verification, signature checks and component store scans show whether the restore worked.

D3-RS at a glance

FieldD3-RS Restore Software
D3FEND IDD3-RS
TacticRestore
Parent techniqueRestore Object (D3-RO): "Restoring an object for an entity to access. This is the broadest class for object restoral."
Siblings under Restore ObjectD3-RC Restore Configuration, D3-RD Restore Database, D3-RDI Restore Disk Image, D3-RF Restore File
Digital artefactSoftware (D3FEND: the technique "restores" Software)
D3FEND referenceCISA: Cybersecurity Incident and Vulnerability Response Playbooks (November 2021)
Typical triggerTampered binaries or modules, malicious add-ins or extensions removed during eradication, security agents uninstalled by an attacker, corrupted system components

Restore Software is distinct from Software Update (D3-SU) in the Harden tactic: updating moves software forward to fix flaws, while restoring returns it to a known-good state. In practice a restore usually ends with the current patched version.

What CISA's playbook asks of a software restore

  • Eradicate first: remove malicious code and account for every persistence mechanism; attackers often maintain several, so allow time to be sure.
  • Rebuild from trusted sources: the playbook's eradication activities include rebuilding or reimaging systems, often from trusted gold images, and its checklist says to reimage affected systems from clean backups ("gold" sources). Hardware may need rebuilding where rootkits are involved.
  • Patch and reset: install updates and patches and reset compromised credentials as part of the same phase.
  • Test before production: test systems, including security controls, before returning them to operational use.

CISA's ransomware guide adds that golden images (preconfigured operating system and application templates) should be kept and regularly updated so systems can be rebuilt quickly, and infrastructure as code templates should be version-controlled and their backups kept offline. See the Server Build and Golden Image Hardening Guide.

ATT&CK techniques D3FEND relates to D3-RS

Relationship type: inferred only. D3FEND relates D3-RS to ATT&CK because these techniques act on the Software artefact, and states that the mappings "are inferred, experimental, and will improve as the knowledge graph grows" (checked November 2026). Restoring software does not prevent these techniques; it removes their effect when the restore source is clean. All IDs were checked against ATT&CK v20.

ATT&CK IDTechniqueWhat the restore must address (NHI Mgmt Group assessment)
T1195.001Compromise Software Dependencies and Development ToolsRebuild from a dependency set confirmed clean; a reinstall that pulls the same poisoned dependency repeats the compromise
T1195.002Compromise Software Supply ChainRestore the version the vendor names as clean, or remove the product until one exists
T1072Software Deployment ToolsCheck the deployment platform and its packages before using it to push restored software
T1554Compromise Host Software BinaryReplace altered binaries from signed packages and verify hashes afterwards
T1137.006Add-insReinstall Office applications and remove unapproved add-ins from user and machine locations
T1505.004IIS ComponentsCompare installed IIS modules with the approved list before and after restoring the web server
T1505.001SQL Stored ProceduresRestore database code from source control, not from the compromised instance
T1574.005, T1574.010Executable Installer File Permissions Weakness; Services File Permissions WeaknessReinstalling restores binaries but not always permissions; check folder and service binary ACLs
T1546.011Application ShimmingRemove unapproved shim databases; reinstalling the target application does not remove them
T1547.008LSASS DriverTreat as a reimage case; component-level repair is not enough when authentication packages were altered
T1542.001, T1542.003System Firmware; BootkitSoftware reinstall does not reach firmware or boot code; reflash firmware from vendor images and rebuild
T1218.014, T1127.001MMC; MSBuildLegitimate tools abused by the attacker; restore them unchanged and control their use instead

The D3FEND page also lists, without IDs on the page, techniques such as Inhibit System Recovery (T1490), Service Stop (T1489) and Software Extensions (T1176), and further IDs including T1056.003 Web Portal Capture, T1497.003 Time Based Checks and T1564.006 Run Virtual Instance. Related guides: MITRE ATT&CK T1195 Supply Chain Compromise, MITRE ATT&CK T1072 Software Deployment Tools, MITRE ATT&CK T1137 Office Application Startup and T1505 Server Software Component Defensive Guide.

Where it is implemented

MechanismWhat it restoresVerification it offers
Reimage from a gold image or infrastructure as codeThe whole operating system and application stackImage version and build record; rebuilt from a template under version control
Windows System File Checker (sfc /scannow)Protected Windows system files, replacing overwritten versions from the local cache or installation sourceScan result for protected files
Windows DISM (/ScanHealth, /CheckHealth, /RestoreHealth)The Windows component store that repairs draw on; /Source names a trusted repair source and /LimitAccess stops DISM using Windows UpdateImage reported healthy, repairable or non-repairable (Microsoft says to discard a non-repairable image)
Linux package managers (for example rpm -V and rpm --reinstall)Files owned by installed packagesrpm -V compares size, digest, permissions, type, owner and group with the package database and flags each difference
Software distribution and endpoint management platformsApplications and agents across many hostsDeployment compliance reports; only as trustworthy as the platform itself
Container and serverless redeploymentWorkloads, by redeploying signed images or packages from the registryImage digest and signature checks at deploy time

Two cautions. First, local repair sources sit on the host being repaired: if the attacker could write to them, use an offline repair source (DISM's /Source with /LimitAccess) or reimage. Second, the package database is itself on the host; rpm -V against a tampered database proves little, so compare with an external baseline where compromise was deep. Products are named to explain mechanisms, not recommended. See the Patch Management Guide and M1045 Code Signing.

Where identity fits

Software restores at scale run through deployment platforms whose service accounts and agents hold administrative rights on every managed host; ATT&CK tracks their abuse as T1072. Before using such a platform to push restored software, confirm its administrator accounts, API tokens and signing keys were not exposed, and rotate them if in doubt. Software that embeds credentials (agents with enrolment tokens, applications with service account passwords in configuration) also needs those secrets reissued after a reinstall. See Securing the patch management tool itself.

Implementation steps

  1. Keep trusted sources ready. Maintain current gold images, an internal repository of signed installers with recorded hashes, and offline repair sources for operating system components.
  2. Preserve evidence. Collect the tampered binaries, modules and add-ins before removal; they show what the attacker changed and help find the same changes elsewhere.
  3. Decide repair or rebuild. Repair in place (SFC, DISM, package reinstall) is acceptable for corruption or limited, well-understood tampering. Rebuild when the attacker had administrative or kernel-level access, altered authentication components, or the scope is unclear, as CISA's playbook favours reimaging from gold sources.
  4. Check the source for supply chain cases. Where the vendor product was the entry point, follow the vendor advisory for the clean version and any additional clean-up steps.
  5. Restore, then update. Install the clean version, apply current patches, and restore approved configuration from D3-RC Restore Configuration sources rather than from the compromised host.
  6. Pilot before fleet-wide rollout. When restoring software across many hosts, push to a small ring first and confirm function and integrity before widening.
  7. Keep rollback ready. Retain the previous clean package and image versions so a faulty restore can itself be reversed.

How to verify the restore

  • Integrity: package verification output shows no unexpected differences; binary signatures validate; hashes match the internal repository records.
  • Inventory: installed software, add-ins, browser extensions and server modules match the approved list for the host role.
  • Security tooling: EDR and logging agents are installed, healthy and reporting before the host returns to production.
  • Function: applications pass functional tests so there is no pressure to bypass controls.
  • Watch period: monitor restored hosts for repeat tampering of the same files or modules; recurrence means persistence was missed.
Metric (example, not a standard)What it showsDirection
Software restores sourced from signed or hash-verified packagesTrust in the restore sourceClose to 100%
Gold images updated within the agreed patch windowReadiness to rebuildRising
Restored hosts passing package or signature verificationRestore effectivenessClose to 100%
Hosts with the same tampering recurring within 30 daysMissed persistence or a tainted sourceFalling
Time to rebuild a critical server from imageRecovery speedStable or falling

Limits and failure patterns

  • Repairing from a local, tainted source: the repair cache or package database on the host was altered, so the repair restores the attacker's version or reports a clean state that is false.
  • Reinstalling over persistence: the application is restored but shims, scheduled tasks, services or add-ins that load it remain. Hunt for persistence before declaring the host clean.
  • Software does not reach firmware: bootkits and firmware implants survive any operating system reinstall.
  • Restoring the vulnerable version: a clean restore of the version that was exploited invites the same intrusion; restore and update together.
  • Trusting a compromised deployment platform: using an attacker-controlled distribution tool to push "clean" software spreads the compromise.
  • Stale gold images: images months behind on patches turn a fast rebuild into a long patch cycle under pressure.

Framework mapping

FrameworkReferenceRelationship to D3-RS
NIST SP 800-53 Rev. 5CP-10 System Recovery and ReconstitutionReturning systems to a known state
NIST SP 800-53 Rev. 5SI-7 Software, Firmware, and Information IntegrityDetecting unauthorised software changes and verifying restored software
NIST SP 800-53 Rev. 5CM-2 Baseline ConfigurationThe approved software baseline the restore returns to
NIST SP 800-53 Rev. 5CM-14 Signed ComponentsRequiring signed software before installation
NIST SP 800-53 Rev. 5SI-2 Flaw RemediationPatching restored software to current versions
NIST CSF 2.0RC.RP-03 The integrity of backups and other restoration assets is verified before using them for restorationChecking images and packages before use

D3FEND cites only the CISA playbook for D3-RS; the NIST rows are NHI Mgmt Group's mapping by intent. For the mitigation view see M1051 Update Software.

Practitioner checklist

  • Keep gold images current and rebuildable from version-controlled templates.
  • Store signed installers with recorded hashes in a protected internal repository.
  • Keep offline repair sources for operating system components.
  • Preserve tampered binaries, modules and add-ins as evidence.
  • Rebuild, not repair, when attackers had administrative or kernel-level access.
  • Follow vendor advisories for supply chain incidents to pick the clean version.
  • Patch immediately after restoring and verify signatures and package integrity.
  • Confirm the deployment platform and its credentials are trustworthy before mass restore.

Standards and references

Related NHI Mgmt Group resources: Server Build and Golden Image Hardening Guide · Patch Management Guide · D3-RF Restore File · MITRE D3FEND Guide

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 12 November 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.