What the CIS Controls Cloud Companion Guide (v8.1) asks customers to do
CIS published the Cloud Companion Guide for CIS Controls v8.1 on December 9, 2024. It explains how to apply the Controls from the consumer or customer side in any cloud environment, whether the workload is IT or OT. If a customer or contract now asks you to map a cloud-hosted estate to the CIS Controls, this guide is the starting point, but it does not give you a separate control set. It is a reading companion to the v8.1 Controls: each top-level Control has a section that explains how to interpret it in the cloud and where cloud considerations differ from a traditional office network. You should read it alongside the v8.1 document, not instead of it.
The answer to "what do we have to do" depends on three things: the provider model (IaaS, PaaS or SaaS), the shared responsibility split in your contracts, and which Safeguards fall to you. A SaaS tenant owns its identities, data and configuration, while the provider runs the platform. An IaaS estate leaves you responsible for most asset, software and configuration Safeguards. Before you read the Controls against your environment, list each cloud service and record, for each Safeguard, whether the provider, you or both carry it. Confirm those splits with the provider's documentation and your contract, because the guide describes considerations, not your specific agreement.
Two areas usually take the most work. The first is Safeguard 4.6, which asks that enterprise assets and software be managed securely, including keeping configuration in version-controlled infrastructure as code and reaching administrative interfaces over secure protocols. In cloud estates, this means checking that deployment templates are versioned and that console, API and SSH access avoids insecure protocols. The second is Control 15, which covers service providers. Safeguard 15.1 asks for an inventory of every known provider with its classification and a named enterprise contact. Your cloud providers belong in that inventory, and the Companion Guide is the place to check how CIS expects that provider relationship to be managed. Use the CIS Benchmarks for your cloud platforms as the next read for measurable configuration settings.
Which cloud Controls need the most attention
Control 4 covers secure configuration of enterprise assets and software. In cloud environments, most configuration is set through consoles, APIs and templates rather than on hardware you own, so the configuration record is often the cloud account itself. Safeguard 4.6 asks that enterprise assets and software be managed securely, and the CIS examples point to keeping configuration in version-controlled infrastructure as code and reaching administrative interfaces over secure protocols such as SSH and HTTPS.
- Version-controlled templates: Keep cloud resource definitions in a repository so changes have an author, a review and a history.
- Secure administrative access: Confirm that management interfaces use SSH or HTTPS and that Telnet and HTTP are disabled unless a documented operational need exists.
- Drift detection: Compare deployed resources against the templates so that console changes do not silently replace approved settings.
- Hardened baselines: Map the account and workload settings to the relevant CIS Benchmarks for your cloud provider.
- Exceptions: Record every insecure protocol that remains and the reason it is needed.
Before an assessment, ask the platform team to show a repository commit that changed a production network setting and the review that approved it.
Control 15 covers service provider management. A cloud platform, managed database, SaaS tool or support partner that holds your data or runs part of your IT operations is a service provider under this Control, even when you never sign a traditional contract with it. Safeguard 15.1 asks for an inventory that lists all known service providers, their classification or classifications, and a designated enterprise contact for each.
- Complete inventory: Include cloud accounts, managed services, SaaS applications and subprocessors that touch sensitive data.
- Classification: Record which providers hold sensitive data or run critical processes, since that drives how much assessment each one receives.
- Named contacts: Assign an internal owner for every provider so that questions about security, incidents and renewals have a destination.
- Shared responsibility: Document which controls the provider operates and which controls your team still owns in each service model.
- Review trigger: Update the inventory when a new cloud service is adopted, not only at annual review.
Ask the team to show the inventory filtered by data sensitivity and confirm that each entry has a named owner.
Control 16 covers application software security across the software life cycle, including software developed in-house, hosted or acquired. Cloud delivery speeds releases through automated pipelines, so security checks have to run inside those pipelines rather than in a separate review at the end. The Cloud Companion Guide links agile, cloud-based development to Control 16, which makes the pipeline a place where evidence is produced.
- Environment separation: Show that development, test and production environments are separated, with distinct accounts or projects where practical.
- Automated code checks: Run static analysis and dependency checks on every merge and record the results.
- Third-party components: Track libraries and container images, and document how vulnerable components are handled.
- Threat modeling: Attach threat models to significant design changes, including new cloud services.
- Pipeline access: Restrict who can change build and deployment definitions, and log those changes.
Ask for a sample of recent pipeline runs that show a failed security check and the fix that followed.
Control 8 covers audit log management. In cloud environments, much of the activity that matters happens inside the provider's control plane, such as account changes, identity events and API calls. Control 8 includes service provider logs, which means you need to collect the logs your cloud provider and other service providers make available to you, not only logs from workloads you operate directly.
- Control plane logs: Enable account-level activity logging and send it to a location your team controls.
- Identity and access events: Capture sign-ins, role changes and permission updates from the cloud identity service.
- Time synchronization: Confirm that logs from cloud and on-premises sources use a consistent time source so events can be correlated.
- Retention: Define how long each log source is kept and who can access it, and test retrieval.
- Provider log review: Name who reviews provider logs and how often, and record the findings.
Before an assessment, check that a sample administrative action in the cloud console appears in your central log store within the period your process states.
CIS publishes Benchmarks for cloud platforms that translate Control 4 into specific settings. For Amazon Web Services, CIS lists current Benchmarks for AWS Foundations, Compute Services, Database Services, End User Compute Services and Storage Services. These Benchmarks give a concrete check for the configuration and identity settings that the cloud Companion Guide asks you to manage. Because Benchmarks change with provider releases, the version you test against matters.
- Foundations first: Start with the AWS Foundations Benchmark for account-level settings, then add service-specific Benchmarks for the services you use.
- Version tracking: Record the Benchmark version used for each scan so results can be compared over time.
- Scanning: Run automated checks against accounts on a schedule and retain the output.
- Remediation ownership: Assign each failed check to a team and track it to closure.
- Scope statement: Document which accounts and services are in scope and which are excluded, with reasons.
Confirm with your platform team which Benchmark versions the scanning tool supports before relying on a score.
How Safeguard 4.6 applies to infrastructure as code and secure admin access
Safeguard 4.6 asks the organization to manage enterprise assets and software securely, and the Cloud Companion Guide applies that request to cloud configuration and administration. The cards below cover the decisions and checks that connect version-controlled configuration, secure administrative access and the evidence an assessor will ask to see. The Assessment Specification defines what is measured, not the tools used to measure it, so the organization must choose and prove its own tooling.
Keep configuration in version control
Safeguard 4.6 points to keeping configuration in version-controlled infrastructure as code. In a cloud account, this means network rules, identity policies, storage settings and compute templates live in repositories rather than only in the console. Check that every production change maps to a reviewed commit, that the repository is the source of truth, and that only approved people can merge to the branch that deploys. Ask who can bypass the repository and how that bypass is logged.
Reach admin interfaces over SSH and HTTPS
Administrative interfaces for servers, hypervisors, container platforms and cloud management consoles should be reached only over encrypted protocols such as SSH and HTTPS. Confirm that management ports are closed to the public internet, that access passes through a bastion, VPN or private endpoint, and that the protocol versions and cipher settings are documented. An assessor will usually test this by attempting a connection from an untrusted network and reviewing the configuration that governs it.
Remove Telnet and HTTP from management paths
Safeguard 4.6 asks that insecure protocols such as Telnet and HTTP be avoided unless operationally essential. Scan cloud load balancers, appliances, legacy images and network devices for listeners on these protocols, and review security group and firewall rules for open ports. Where a legacy system still needs one of them, record the business reason and the compensating control in the exception register described in the next card.
Document every secure configuration exception
Any use of an insecure protocol or a deviation from the hardened baseline needs a written exception. A useful record names the asset, the protocol or setting, the business owner, the reason it is operationally essential, the compensating control, the review date and the approver. Assessors commonly sample these records and compare them with live configuration. Confirm who reviews exceptions and how an expired exception is closed or renewed.
Evidence the configuration history
Version control gives the audit trail for infrastructure changes. Keep commit history, pull request approvals, pipeline run logs and the change tickets they reference, and make sure they are retained for the period the organization sets in its policy. Ask whether the pipeline records which commit was applied to which account and when, because an assessor will often ask you to trace a sampled production setting back to an approved change.
Check for configuration drift
Drift occurs when a live cloud resource no longer matches its code, often after a console change during an incident. Schedule automated comparisons between the deployed state and the repository, alert on differences and route each finding to a remediation or exception record. Confirm which tool performs the comparison, how often it runs and who receives the alerts. The Assessment Specification does not prescribe the tool, so the organization must show that its chosen method runs and that findings are acted on.
Which CIS AWS Benchmarks map to which cloud footprint
This table lists the five CIS AWS Benchmarks supplied for October 2026, with the version of each, the AWS workload it covers and how it connects to Control 4 secure configuration. Use it to decide which Benchmarks apply to your account footprint, then confirm the current version on the CIS Benchmarks page before you start hardening.
| Benchmark | Version | Workload it covers | Control 4 link |
|---|---|---|---|
| AWS Foundations | 7.1.0 | Account-level baseline: identity, logging, monitoring and networking settings that apply across the whole AWS account | Sets the secure configuration baseline for enterprise assets and software under Control 4 in every account |
| AWS Compute Services | 2.0.0 | Compute resources such as virtual machines and container and serverless compute services | Hardens the instances and services that run your workloads, including administrative interface settings |
| AWS Database Services | 2.0.0 | Managed database services, including their configuration, access and encryption settings | Covers secure configuration of data stores that hold sensitive data, supporting Safeguard 4.6 expectations |
| AWS End User Compute Services | 1.2.0 | End user computing services such as virtual desktops, relevant if staff or contractors use cloud-hosted workstations | Applies secure configuration to user-facing cloud endpoints, which Control 4 treats as enterprise assets |
| AWS Storage Services | 1.0.0 | Object and file storage services, including public access and encryption settings | Configures storage so that data is not exposed through insecure defaults, aligned with Control 4 hardening |
How to run a cloud-focused CIS Controls assessment
This sequence turns the CIS Controls v8.1 into a cloud-specific assessment: it fixes the Implementation Group and scope first, then maps Safeguards to each cloud platform, gathers configuration and provider evidence, and ends with a prioritized roadmap and, where required, a handoff to an independent assessor.
Confirm the applicable Implementation Group
The organization's security lead, with input from risk and business owners, decides which Implementation Group applies: IG1 as essential cyber hygiene, IG2 for greater operational complexity or sensitive information, or IG3 for regulated, sensitive assets facing sophisticated attacks. Every Safeguard in scope is then tied to that group, so the assessment has a fixed target. The output is a short written statement naming the group, the reasons for it, and the Safeguards it includes.
Scope cloud accounts and services
Platform engineering and cloud owners list every cloud account, subscription, project, region and managed service that stores, processes or transmits in-scope data, including any IT or OT workloads that run in the cloud. Each item is tagged with an owner and linked to the business process it supports. The scope is complete when every account in the cloud inventory is either in scope with a named owner or excluded with a written reason.
Read the companion guide alongside v8.1
The assessment lead reads the Cloud Companion Guide section for each top-level Control next to the v8.1 Controls document, noting where the guide explains cloud-specific interpretation or considerations that differ from a traditional office network. Findings are recorded per Control, not per Safeguard, so the team can see which cloud assumptions change the wording of a requirement. The output is an annotated copy of the Controls with a list of cloud interpretations the team has adopted and why.
Map Safeguards to each cloud platform
Engineers take each in-scope Safeguard and identify the native cloud service, configuration setting or managed control that satisfies it on each platform in scope. The Benchmarks for the relevant provider, such as the CIS AWS Benchmarks for foundations, compute, database, end user compute and storage, can be used as a reference for configuration checks. Where no native control exists, the gap is recorded with the compensating process. The output is a platform-by-Safeguard matrix that names the responsible owner for each row.
Collect configuration and provider evidence
Security engineers gather evidence that matches the measures in the CIS Controls Assessment Specification for each Safeguard. For Safeguard 4.6 this includes infrastructure as code kept in version control, evidence that administrative interfaces use secure protocols, and proof that insecure protocols are disabled unless operationally essential. For Safeguard 15.1 this includes the service provider inventory with classifications and named enterprise contacts. The output is an evidence folder indexed by Safeguard, with the date and source of each artifact.
Run the improvement assessment and build the roadmap
The assessment team scores each Safeguard as implemented, partially implemented or not implemented, using the Assessment Specification's measures as the test. Gaps are ranked by risk and by how much they affect the Implementation Group target, and each one receives an owner and a dependency list. The output is a prioritized improvement roadmap that the organization owns and can update as cloud services change. Ask for the scoring method and roadmap format in writing before work starts.
Hand evidence to an independent assessor
If a customer, contract or regulator requires a formal assessment or attestation, the organization gives the evidence package and scope statement to an independent assessor or certification body, which decides what to test and issues any report. The implementation partner can prepare the systems and evidence for that review but does not issue reports or certificates. The step is complete when the assessor confirms receipt and the organization records the assessor's scope and any findings it must address.
What must be in place before a cloud Controls review
A cloud Controls review tests whether the organization can show what it runs, who runs it and how it is changed. The six items below connect the cloud Companion Guide to the Safeguards that an assessor will ask about first, and each one can be checked before the review starts.
Current cloud asset and account inventory
Start with a list of every cloud account, subscription, project and workload in scope, with its owner and environment (production, test or development). The Companion Guide applies Control 1 and Control 2 to cloud resources, so ephemeral instances, managed services and serverless functions must appear too. Check the inventory against the provider's own account listing and look for accounts that exist in billing but not in the register. A stale inventory is the most common reason a review stalls at its first step.
Documented provider responsibilities
Write down which security tasks the cloud provider performs and which the organization keeps, such as identity configuration, encryption key management, network rules and patching of managed services. The shared responsibility split differs by service model, so a single statement for the whole estate usually fails. Confirm the split against the provider's published documentation and record the version you relied on. Assessors look for this boundary before they test any Safeguard that depends on it.
Infrastructure as code repositories with history
Safeguard 4.6 expects enterprise configuration to be kept in version-controlled infrastructure as code. Confirm that each production environment is built from templates in a repository, that changes go through reviewed commits, and that the history shows who changed what and when. Check for manual console edits that never reached the repository, because they break the link between the approved configuration and the running one. Administrative access should reach these systems only over protocols such as SSH and HTTPS.
Provider inventory under Safeguard 15.1
Safeguard 15.1 asks for an inventory of all known service providers, each with its classification and a designated enterprise contact. For a cloud review, list every cloud provider, SaaS platform and managed service that holds sensitive data or supports critical processes, not only the main hosting vendor. Check that each entry names a business owner who can answer for the contract and the security terms. Gaps usually appear around smaller managed services adopted by individual teams.
Logging of provider and administrator activity
Confirm that control plane activity, such as account changes, role assignments, network rule edits and key use, is recorded and sent to a store the administrators cannot quietly alter. Ask whether the logs cover the provider's own audit trail, the organization's accounts and the service provider accounts used for management. Test retrieval by pulling a sample event for a known change from last month. If the logs cannot be retrieved on demand, the review cannot confirm who did what.
Named owners for each Safeguard
Assign one accountable person to each in-scope Safeguard, recorded with a role, a backup and the evidence location. Without named owners, a cloud review tends to produce evidence that nobody can explain or refresh. Check that owners include both platform and security roles, and that the list is reviewed when staff move on. The organization owns these controls; an implementation partner can help prepare the register, but it cannot take over accountability for the Safeguards.
Questions cloud teams ask about the CIS Controls companion guide
Is the CIS Controls Cloud Companion Guide a separate certification?
No. The Cloud Companion Guide is guidance that explains how to read and apply the CIS Controls v8.1 in cloud environments, and it does not create a certification or a compliance report. CIS states that the Controls do not replace any regulatory, compliance or authorization scheme. If a contract or regulator asks for evidence, the organization needs to confirm with that requester which assessment or attestation they accept, and which independent auditor or assessor issues it. Read the guide alongside the CIS Controls v8.1 document, since it adds cloud-specific considerations to each top-level Control rather than replacing the Safeguards.
Can an implementation partner prepare our cloud environment for CIS Controls work?
Yes, a partner can prepare systems, configurations and evidence, but the organization still owns its controls. In practice, an implementation partner can help build infrastructure as code, harden cloud accounts, centralize logging and document processes so the organization can show how each Safeguard is met. Accountability for decisions, risk acceptance and ongoing operation stays with the organization. Only an independent auditor, assessor or certification body issues reports or certificates. DeployOpen plans, implements and documents self-hosted open-source platforms inside the client's own cloud accounts, and it prepares evidence for assessments without issuing audit results. Confirm the division of responsibilities in the written scope before work starts.
What drives the cost of applying the companion guide to our cloud?
No public figure for a cloud engagement is supplied here, so the useful question is what drives the cost. The main drivers are the number of cloud accounts, subscriptions and regions in scope, the number of workloads and IaC repositories that must be brought under version control, the state of existing configuration and logging, and how many service providers and cloud-hosted applications must be inventoried and assessed. One-time costs usually include the baseline assessment, hardening, pipeline and logging changes and documentation. Recurring costs include monitoring, patching, periodic review and any retainer for ongoing support. When requesting a quote, ask for the scope in writing, the deliverables, who performs each task, which items are one-time versus recurring, and what the organization must supply.
How long does a cloud CIS Controls improvement project take?
Duration depends on the scope, so confirm the timeline with the partner against a written statement of work. The main variables are the number of cloud environments, the maturity of existing infrastructure as code, the number of Safeguards being addressed, and how quickly the organization can review and approve changes. The companion guide is designed to support an improvement assessment and roadmap, so a sensible first phase is a baseline assessment that identifies gaps, followed by prioritized remediation. Ask the partner to break the timeline into phases with named deliverables, state which dependencies sit with your teams (access, change approvals, owners for each system), and specify how progress will be reported.
Does the companion guide cover operational technology running in the cloud?
Yes. CIS describes the Cloud Companion Guide as covering how to apply CIS Controls v8.1 in any cloud environment from the consumer or customer perspective, for both IT and OT cloud environments. The guide does not replace the Controls for OT. Organizations running industrial or operational workloads in the cloud should check how the Safeguards apply to those systems, especially where availability, safety or vendor-managed equipment limit what can be changed. Confirm with your OT engineers and your partner which assets are in scope, how segmentation and remote access are handled, and which Safeguards need an adapted approach before any hardening begins.
How do we map cloud results to other frameworks using the Navigator?
Use the CIS Controls Navigator for v8.1 to start. It offers CIS mappings from the Safeguards to 30 frameworks and versions as of October 2026, including NIST CSF 2.0, NIST SP 800-53 Rev. 5, ISO/IEC 27001:2022, SOC 2, PCI DSS v4.0, DORA and NIS2. Build a single evidence register keyed by Safeguard, then tag each item with the mapped framework requirements. Treat the Navigator as CIS's published mapping rather than a substitute for your own control analysis, and have your assessor confirm how each requirement is judged. Where a framework is not in the Navigator, document the mapping yourself and mark it as your interpretation so the auditor can review it.
How DeployOpen prepares cloud estates for a CIS Controls review
DeployOpen is a Chainscore Labs company that plans, implements, hardens, documents and supports self-hosted open-source platforms inside the client's own cloud accounts. For CIS Controls Cloud Companion Guide (v8.1), we prepare systems and evidence for an assessment, while the organization owns its controls and an independent assessor reaches any conclusion.
How we scope
- Engagements start with a written scope that names the cloud accounts, workloads and Safeguards in question.
- The scope states which Controls and Safeguards the cloud estate covers, including Safeguard 4.6 and Control 15 where they apply.
- Scope, deliverables and timelines are agreed in writing before any work begins, and a quote follows from that scope.
How we build
- We configure self-hosted open-source platforms in your own cloud environment, using infrastructure as code kept under version control.
- Administrative interfaces are reached over secure protocols, and insecure protocols are disabled unless the organization documents an operational need.
- We test the built configuration against the relevant CIS Benchmarks for your cloud provider and record the results for review.
How we hand over
- Deliverables include configuration documentation, asset and service provider inventories, and a record of each Safeguard's implementation status.
- Evidence is packaged so your internal team and independent assessor can trace each claim to a configuration, log or test result.
- Handover sessions walk your staff through the environment so they can operate and update the controls after the engagement closes.
How we support
- Ongoing support, upgrades and continuous testing are available as retainers after the initial handover is complete.
- Retainers can keep Benchmark checks, configuration baselines and evidence collection current as cloud services and versions change.
- Confirm with us in writing which support tasks are included, how changes to scope are handled, and who owns each evidence artifact.

