Most configuration weaknesses survive because nobody was ever required to adopt a baseline, measure against it or put it back when it drifted. This template gives you a numbered policy that mandates approved secure baselines for every asset class, sets how they are chosen and changed, requires automated drift detection and remediation, and routes every deviation through a controlled exception process. It follows the four phases of security-focused configuration management in NIST SP 800-128 and maps to CIS Control 4 and NIST SP 800-53 CM-2, CM-6 and CM-7. It is written for security, infrastructure and risk leads who need a policy that engineering teams can follow and auditors can test. Exceptions themselves are handled in the companion Hardening Exceptions and Compensating Controls Template.
How to use this template
- Replace text in square brackets, such as [Organisation], [CISO], [Configuration Control Board] and [30 days], with your own names and values. Bracketed values are illustrative starting points, not regulatory requirements.
- "Must" is mandatory; "should" is recommended practice where a documented deviation is acceptable.
- Keep this policy short and stable. Name the actual baselines, versions and tooling (Group Policy, Intune, Ansible, Terraform policy checks) in a supporting Secure Configuration Standard that can change without re-approving the policy.
- Use this policy with your Change Management Policy Template (baseline changes are changes) and Asset Management Policy Template (you cannot baseline what you have not inventoried).
- Prefer adopting a published, tested baseline over writing your own. Microsoft states this for its Windows security baselines, and the same reasoning applies to CIS Benchmarks and vendor hardening guides.
Choosing baseline sources
| Source | What it is | Best fit | Watch out for |
|---|---|---|---|
| CIS Benchmarks | Consensus-based, prescriptive configuration recommendations for operating systems, cloud, network devices, databases and applications, with CIS-CAT Pro for assessment | Most organisations wanting a vendor-neutral, widely recognised reference | Higher profiles (for example Level 2) can break applications; pilot before enforcing |
| DISA STIGs (and CIS STIG Benchmarks) | US Department of Defense configuration requirements; CIS publishes STIG-based benchmarks | Defence supply chain and contracts that require STIG compliance | Very restrictive; many findings need documented acceptance |
| Vendor baselines (Microsoft security baselines, cloud provider guidance) | Vendor-recommended settings, delivered for Windows through the Security Compliance Toolkit, MDM baselines and Intune | Platforms where the vendor tracks new settings fastest | Cover one vendor's products only; check version currency |
| Government guidance (NCSC device security guidance, CISA SCuBA, NIST National Checklist Program) | Platform guides and configuration packs; the NIST repository indexes checklists by product | Public sector and organisations aligning to national guidance | Coverage varies by platform |
| Internal baseline | Your own standard derived from one of the above with documented deviations | Always, as the controlled version you actually enforce | Must record which source and version it derives from |
1. Purpose
This policy requires that every technology asset owned, managed or operated by or for [Organisation] is deployed and kept in an approved secure configuration, so that default settings, unnecessary functionality and unmanaged drift do not create exploitable weaknesses. It supports [Organisation]'s [Information Security Policy] and implements security-focused configuration management as described in NIST SP 800-128.
2. Scope
This policy applies to:
- Assets: end-user devices including laptops and mobile devices, servers, virtual machines and container images, network and security devices, IoT and other non-computing devices, operating systems, databases, middleware and business applications, consistent with the scope of CIS Control 4.
- Cloud and SaaS: cloud accounts, subscriptions and projects, managed services, Kubernetes clusters, and the security settings of SaaS tenants (for example identity, sharing and logging settings).
- Code-defined configuration: infrastructure as code, golden images, configuration management manifests and deployment pipelines that set configuration.
- People: all staff, contractors and third parties who build, configure or administer in-scope assets.
Operational technology follows this policy unless the [OT Security Standard] sets a stricter or safety-driven alternative.
3. Roles and responsibilities
| Role | Responsibilities |
|---|---|
| [CISO] | Owns this policy; approves baseline sources and risk-rated exceptions above [high]; reports compliance to [Risk Committee] |
| [Configuration Control Board] | Reviews and approves new baselines and baseline changes, including security impact analysis, in line with the NIST SP 800-128 CCB model (a less formal board may be used for low-risk systems) |
| [Security Engineering] | Maintains the Secure Configuration Standard, tailors baselines, publishes scan content and drift rules |
| Platform and system owners | Apply the baseline, remediate drift within target times, request exceptions, keep build artefacts (images, templates) compliant |
| [Security Operations] | Monitors drift and unauthorised change alerts, escalates suspected tampering as a security incident |
| [Internal Audit] | Tests baseline coverage, scan evidence and exception handling independently |
4. Policy statements
4.1 Establishing baselines
- [Security Engineering] must define and maintain an approved secure baseline for each asset class in scope, derived from a recognised source (see the table above) and recording the source name and version.
- Each baseline must state which settings are tailored or not applied and why. Tailoring that weakens a control must be approved by the [Configuration Control Board].
- Baselines must address, at minimum: removal or disabling of unnecessary services, ports, protocols and software (least functionality); default and built-in accounts; authentication and session lock settings; host firewall; logging and audit settings; encryption at rest and in transit; and trusted DNS.
- Baselines must be reviewed at least [annually] and when the vendor releases a new major version, a new baseline version is published, or a relevant threat or incident arises.
4.2 Building and deploying
- New assets must be built from an approved, hardened image, template or configuration manifest. Manual builds are permitted only where no automated route exists and must be checked against the baseline before entering production.
- Golden images and infrastructure-as-code modules must be scanned against the baseline in the pipeline, and builds that fail [critical] checks must not be released.
- No asset may enter production with vendor default passwords, sample content or unused default accounts enabled.
4.3 Default accounts and machine credentials
- Default and built-in accounts must be disabled, renamed where the platform supports it, or have their credentials replaced before production use.
- Service accounts, API keys and tokens used by configuration management, scanning and deployment tools must be individually owned, least privileged, stored in [approved secrets manager] and rotated per the [NHI Policy]. These tools can change every system they manage, so their identities are treated as privileged.
- Local administrator passwords must be unique per device and managed centrally.
4.4 Controlling change
- Changes to an approved baseline must follow the [Change Management Policy], including security impact analysis and testing, before implementation.
- Configuration must be enforced from a central source of truth (for example Group Policy, MDM, configuration management or IaC) wherever the platform supports it, so that local changes are overwritten or flagged.
4.5 Monitoring and remediation of drift
- All in-scope assets must be assessed against their baseline automatically at least [weekly], and continuously where the tooling supports it.
- Deviations must be remediated within: [7 days] for internet-facing and critical assets with a [critical] finding; [30 days] for other [critical] or [high] findings; [90 days] for the remainder, unless an approved exception exists.
- Unauthorised changes to security-relevant settings (for example disabled logging, endpoint protection or firewall) must raise an alert to [Security Operations] and be treated as a potential security incident.
- Assets that cannot be assessed (agent missing, scan failing) must be reported as non-compliant, not excluded.
5. Exceptions
- Any setting that cannot meet the baseline must have an approved exception recorded in the [exception register] before or, for emergencies, within [5 working days] of the deviation.
- Each exception must name an owner, the business reason, compensating controls, a risk rating and an expiry date of no more than [12 months].
- Exceptions are managed through the Hardening Exceptions and Compensating Controls Template. Expired exceptions count as non-compliance.
6. Compliance and monitoring
[Security Engineering] reports the following to [Risk Committee] [monthly]:
| Metric | Definition | Target |
|---|---|---|
| Baseline coverage | Percentage of inventoried assets with an assigned baseline and a successful assessment in the last [7 days] | [98%] |
| Configuration compliance score | Percentage of baseline checks passing, by asset class | [95%] for critical assets |
| Drift remediation time | Median and 90th percentile days from detection to fix, by severity | Within section 4.5 targets |
| Build compliance | Percentage of images and IaC releases passing baseline checks first time | Rising trend |
| Exception health | Open, expired and overdue exceptions | Zero expired |
| Default credentials found | Assets with default or shared vendor credentials detected | Zero |
Breaches of this policy may lead to removal of the asset from the network and disciplinary action under [HR policy].
7. Review and approval
This policy is reviewed at least [annually] by [CISO] and approved by [Executive Committee]. Version: [1.0]. Effective date: [date]. Next review: [date].
Implementation guidance
Roll the policy out in the order NIST SP 800-128 describes: plan, identify and implement baselines, control change, then monitor.
- Inventory and classify: confirm asset classes from your inventory and pick one baseline source per class. Start with endpoints, domain controllers and internet-facing servers.
- Assess before enforcing: scan a representative set against the chosen baseline in report-only mode. Expect many findings on first run; record which settings break applications.
- Tailor and approve: remove or relax settings that break business services, record each as tailoring or an exception, and have the [Configuration Control Board] approve version 1 of the internal baseline.
- Pilot in rings: enforce on IT and security staff devices, then a pilot business group, then everyone. Keep the previous Group Policy, MDM profile or manifest version so you can roll back within one policy refresh cycle.
- Bake into builds: move the baseline into golden images and IaC modules so new assets start compliant. See the Server Build and Golden Image Hardening Guide and the Infrastructure as Code Security Guide.
- Turn on drift detection: alert on security-relevant deviations and feed them to the SOC. The Configuration Drift and Baseline Enforcement Guide covers tooling.
- Report and review: start monthly metrics, then review baselines when new vendor versions ship.
Small and large organisations
- Small: adopt the vendor or CIS baseline closest to Level 1, enforce it through your MDM or a single Group Policy set, and let one named person act as the change board. Use the Endpoint Hardening Checklist to start.
- Large: run a formal board, version baselines in source control, enforce through pipelines, and split metrics by business unit and asset class.
Common failure patterns
- Baseline adopted, never measured: a document exists but no scan runs, so compliance is assumed.
- Golden image drift: images are hardened once, then patched and changed without re-scanning, so every new build inherits weaknesses.
- Unassessable assets excluded: agents fail silently and those hosts drop out of the score instead of counting against it.
- Exceptions without expiry: temporary relaxations become permanent and invisible.
- Over-trusted tooling: configuration tools run with domain-wide or organisation-wide rights and long-lived keys. See the Service Account Security Guide.
Framework mapping
| Policy section | NIST SP 800-53 Rev. 5 | NIST CSF 2.0 | CIS Controls v8.1 | ISO/IEC 27001:2022 |
|---|---|---|---|---|
| 4.1 Establishing baselines | CM-2 Baseline Configuration, CM-6 Configuration Settings, CM-7 Least Functionality | PR.PS-01 | Safeguard 4.1, 4.8 | Annex A 8.9 Configuration management |
| 4.2 Building and deploying | CM-2(2) Automation Support for Accuracy and Currency | PR.PS-01 | Safeguard 4.1 | Annex A 8.9 |
| 4.3 Default accounts | CM-6 | PR.PS-01 | Safeguard 4.7 | Annex A 8.9 |
| 4.4 Controlling change | CM-3 Configuration Change Control | PR.PS-01 | Safeguard 4.1 | Annex A 8.32 Change management |
| 4.5 Drift monitoring | CM-6(1), CM-6(2) Respond to Unauthorized Changes, CA-7 Continuous Monitoring, SI-7 | DE.CM-09 | Safeguard 4.1 | Annex A 8.9 |
| 2 Scope (inventory) | CM-8 System Component Inventory | ID.AM-08 | Control 1, Control 2 | Annex A 5.9 |
Sources: NIST SP 800-128, NIST SP 800-53 Rev. 5, CIS Control 4, CIS Benchmarks, Microsoft Windows security baselines, NIST National Checklist Program and NCSC device security guidance.
This template is provided for guidance and must be adapted to your organisation. It is not legal advice. Related NHI Mgmt Group resources: CIS Control 4 Secure Configuration of Enterprise Assets and Software Guide · Hardening Exceptions and Compensating Controls Template · CISA SCuBA Baselines Guide · NHI Policy Template