Join our Newsletter — 33% off our NHI Course
Home› Guides› Secure by Design for UK Government Guide: Principles,…
Guide Regulations & Frameworks

Secure by Design for UK Government Guide: Principles, Roles, Evidence and Rollout

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 27 October 2026 14 min read
On this page

UK government digital services are still too often secured at the end of delivery, when design choices that create risk are already expensive to change. Secure by Design is the UK government framework, published by the Government Cyber Unit on security.gov.uk, that tells delivery teams and security professionals how to build cyber security into digital services and technical infrastructure from the start and keep it there for the life of the service. This guide explains what the framework contains, who it applies to, how the 10 principles, the activities and the Self Assessment Tracker fit together, and where identity and non-human identity controls belong. It is written for delivery managers, technical architects, security leads and suppliers working on central government services. It is not the US CISA Secure by Design pledge, which targets software manufacturers; for that policy strand see the Software Liability and Secure by Design Policy Guide.

Key takeaways

  • It is a requirement, not optional guidance, for central government. The implementation page states that departments and arm's length bodies (ALBs) must incorporate effective security practices and meet the Secure by Design policy, and Service Standard point 9 says service teams must follow the principles.
  • Ten principles carry the substance. They run from creating responsibility for cyber security risk to making changes securely, and every activity on security.gov.uk is written to help teams meet one or more of them.
  • Evidence is a self-assessment, not a certificate. Teams complete the Secure by Design Self Assessment Tracker per delivery phase (Discovery, Alpha, Private beta, Public beta or live), and an internal assurance team assesses the evidence independently.
  • Accountability sits with named people. Chief digital information officers are accountable for adoption, and the senior responsible owner (SRO) and service owner sign off the tracker before it goes to assurance.
  • Identity is where many principles become concrete. Least privilege for service accounts, short-lived credentials for pipelines and integrations, and logging of machine activity are the practical way to meet the attack surface, defence in depth and detect and respond principles.

What it is and who it applies to

Secure by Design is made up of a policy, a set of principles and a library of activities, supported by case studies and a self-assessment tool. The About page describes it as a core requirement of the government Cyber Security Standard and says it was developed by the Government Cyber Unit with a cross-government working group, the Government Security Group, the National Cyber Security Centre (NCSC) and industry experts. It also states that Secure by Design is not an assurance process in itself: delivery teams complete a self-assessment as evidence that they meet the principles.

GroupWho is in it (as stated on the implementation page)What it means in practice
Group 1Ministerial departments, ALBs managing government Critical National Infrastructure, and organisations managing priority government servicesEarliest adopters; implementation schedules agreed directly with the Government Cyber Unit Secure by Design team
Group 2All remaining ALBs and other central government organisationsSame policy, schedule agreed later; encouraged to start as soon as possible
SuppliersCommercial contracts where the buyer includes Secure by Design requirementsRequirements apply only where the contract says so; expect them in statements of requirement and assurance evidence requests
DefenceMinistry of Defence programmes and defence industryThe MoD runs its own Secure by Design version, mapped to its environment and project life cycle

Scope of work: the requirement applies to digital services and technical infrastructure, and the implementation page says it is particularly important for new services and for services undergoing significant change. Existing services in steady state are not explicitly scoped in, but principle 9 (embed continuous assurance) and principle 10 (make changes securely) mean that any significant change pulls a live service into the process.

Spend and assurance gates: from 1 April 2026 the Digital Assurance Playbook replaced the digital and technology spend control approval process, and the playbook asks teams to check they follow their organisation's Secure by Design approach. In practice this makes Secure by Design evidence part of the funding conversation rather than a separate security exercise.

Status note: the security.gov.uk pages are updated regularly (the implementation and About pages were updated in 2026) and some related content requires a government sign-in. Check the live pages for your organisation's group and schedule rather than relying on a fixed date.

Structure and key requirements

The 10 principles are the requirement that service teams are measured against. The activities are recommended practical guidance; teams can meet a principle in other ways if they can evidence it.

#PrincipleWhat it asks forExample activities on security.gov.uk
1Create responsibility for cyber security riskSenior risk owners accountable for security risk through the service lifeAgreeing roles and responsibilities; identifying security resources; considering security within the business case
2Source secure technology productsAssess third-party platforms, software and code, reduce the risks and share findings with suppliersManaging third-party product security risks
3Adopt a risk-driven approachA defined risk appetite and continuous assessment of threatsWorking out the project's security risk appetite; sourcing a threat assessment; performing a security risk assessment
4Design usable security controlsUser research so security processes are practical and not worked aroundUnderstanding business objectives and user needs
5Build in detect and respond securityLogging, monitoring, alerting and response, tested regularlyManaging observability
6Design flexible architecturesComponents that can take new controls as threats changePerforming threat modelling; agreeing a security controls set
7Minimise the attack surfaceOnly the capabilities, software, data and hardware the service needsDocumenting service assets; retiring service components securely
8Defend in depthLayered controls so one failure does not lead to full compromiseResponding to and mitigating security risks
9Embed continuous assuranceOngoing evidence that controls work, at delivery and in operationAssessing the effectiveness of security controls; discovering vulnerabilities; implementing a vulnerability management process
10Make changes securelySecurity impact considered for every design, development and deployment changeEvaluating the security impact of changes; tracking Secure by Design progress

Choosing a controls set

The activity "Agreeing a security controls set for your service" asks the technical architect and the organisation's security team to shortlist frameworks from what needs protecting, combine them into one agreed controls set, share it with the delivery team and revisit it as the service changes. The frameworks it lists as examples include the NCSC Cyber Assessment Framework (CAF), the NIST Cybersecurity Framework and SP 800-53, ISO/IEC 27002:2022, the CIS Critical Security Controls, Cyber Essentials, the NCSC 10 Steps and Cloud Security Principles, the NCSC zero trust architecture principles, the Cloud Security Alliance Cloud Controls Matrix, MITRE ATT&CK Mitigations, the ISF Standard of Good Practice and the OWASP Top 10. An example taxonomy spreadsheet maps framework controls to CAF 4.0.

  • Small service or SaaS adoption: start from the NCSC Cloud Security Principles or the CSA Cloud Controls Matrix plus Cyber Essentials, as the activity suggests for SaaS. See the Cyber Essentials Guide.
  • Large or CNI-adjacent service: anchor on CAF outcomes, which many departments are already assessed against, and use SP 800-53 or ISO/IEC 27002 for control detail. See the NCSC Cyber Assessment Framework (CAF) Guide.
  • Web service: add the OWASP Top 10 as the activity recommends, and use OWASP ASVS for testable requirements.

Identity and NHI requirements

The Secure by Design pages do not set identity-specific rules; identity controls come in through the controls set you agree (for example CAF objective B2 on identity and access control, or SP 800-53 AC and IA families). In real services the principles are most often broken through identity: a pipeline token with tenant-wide rights, a forgotten integration account, or an API key in a repository. The table below is NHI Mgmt Group's own indicative reading of how the principles apply to identities, not official Government Cyber Unit guidance.

PrincipleIdentity and NHI practice that evidences it
1 Responsibility for riskEvery service account, workload identity and API key has a named human owner recorded in the service asset register. See the NHI Ownership Guide.
2 Secure productsSupplier questions cover how the product authenticates to your estate, what scopes its integrations need and whether it supports federated or short-lived credentials instead of static keys.
5 Detect and respondMachine sign-ins and token use are logged and baselined, not only human logins. See the Detection Engineering for Non-Human Identities Guide.
7 Minimise attack surfaceNo standing admin for automation; secrets held in a vault, not configuration files; unused identities removed when components are retired. See the Secrets Management Guide.
8 Defend in depthLeast privilege scopes so a stolen deployment credential cannot reach production data; separate identities per environment. See the Least Privilege Techniques Guide.
10 Change securelyNew integrations and permission changes go through the same security impact evaluation as code changes. See the NHI Offboarding and Decommissioning Guide for the retirement side.

A practical test for the Alpha phase: can the team list every non-human identity the service will create, what it can reach and who owns it? If not, the "Documenting service assets" activity is incomplete. The Service Account Inventory Template gives a starting structure.

Implementation steps

The implementation page includes a preparation checklist for organisations to assess current compliance and identify gaps. A workable sequence for a department or ALB looks like this:

  1. Confirm your group and schedule. Contact the Government Cyber Unit Secure by Design team if your implementation schedule has not been agreed, and record the CDIO as accountable owner.
  2. Run the preparation checklist. Compare existing delivery governance, risk management and assurance processes with the policy, and list gaps.
  3. Agree roles per project. The SRO, service owner and product and delivery managers agree where security responsibilities sit, consulting security professionals. Use the example RACI matrix linked from the "Agreeing roles and responsibilities" activity; it defines one accountable person (or two where ownership is shared) per task.
  4. Pilot on one or two new services. Pick services entering Discovery or Alpha, complete the activities relevant to that phase and fill in the tracker. Collect feedback on effort and usefulness.
  5. Agree an organisational controls set and taxonomy. Reuse it across projects so teams do not each select frameworks from scratch; map it to CAF if you are assessed under GovAssure.
  6. Connect to spend and assurance gates. Make the tracker output an input to Digital Assurance Playbook reviews and to phase gates (alpha and beta assessments under the Service Standard).
  7. Scale and embed. Extend to all new services and significant changes, include security tasks in job descriptions as the roles activity suggests, and review the approach annually.

Rollback and correction: if evidence no longer supports a "Yes" answer, the tracker guidance says the answer must change to "No" and affected people must be told. Treat this as the correction mechanism: a downgrade is a signal to fix, not a failure to hide.

Evidence and assessment

The Secure by Design Self Assessment Tracker is an Excel workbook with a project profile sheet and one sheet per phase: Discovery, Alpha, Private beta, and Public beta or live. Non-agile projects map their stages onto these phases.

ElementWhat the guidance saysPractitioner note
AnswersYes, No or N/A per question, linked to recommended activities; N/A needs an explanationWrite N/A reasons a reviewer can test, such as "no personal data processed in Discovery"
EvidenceNotes on how each requirement was met or where outputs are held; links optional and access-restrictedPoint to versioned artefacts (threat model, risk assessment, controls set) rather than prose
Confidence profileLow, medium or high per phase, starting at Low; answers are weightedA high profile does not guarantee security, and the tracker is not a risk register
Review and sign-offDelivery managers coordinate; technical and security staff (including CISO and CTO) confirm; SRO and service owner sign off; the internal assurance team assesses independentlyBook assurance time before phase gates, not after
HandlingThe tracker is a restricted assetIt describes weaknesses: store and share it accordingly

Measuring adoption

  • Coverage: percentage of new services and significant changes with a current tracker for their phase.
  • Confidence trend: share of services reaching medium or high confidence by Private beta.
  • Finding closure: median days to close risks raised in security risk assessments and third-party product reviews.
  • Identity hygiene: percentage of non-human identities in scope with a named owner and no standing administrative rights.
  • Detection readiness: share of services whose logging and alerting have been tested before go-live, as principle 5 requires.

Mapping to other frameworks

No official crosswalk from the 10 principles to NIST or ISO was found on security.gov.uk at the time of writing (the official example taxonomy maps controls to CAF 4.0). The table below is NHI Mgmt Group's own indicative mapping, intended to help organisations reuse existing control evidence. Validate it against your controls set.

PrincipleNIST CSF 2.0 (indicative)NIST SP 800-53 Rev. 5 (indicative)
1 Responsibility for riskGV.RR Roles, Responsibilities, and AuthoritiesPM-9 Risk Management Strategy
2 Secure productsGV.SC Cybersecurity Supply Chain Risk ManagementSA-9 External System Services; SR-3 Supply Chain Controls and Processes; SR-5 Acquisition Strategies, Tools, and Methods
3 Risk-driven approachGV.RM Risk Management Strategy; ID.RA Risk AssessmentRA-3 Risk Assessment
5 Detect and respondDE.CM Continuous Monitoring; RS.MA Incident ManagementAU-6 Audit Record Review, Analysis, and Reporting; IR-4 Incident Handling
6 and 8 Flexible architecture, defence in depthPR.PS Platform Security; PR.IR Technology Infrastructure ResilienceSA-8 Security and Privacy Engineering Principles; SC-7 Boundary Protection
7 Minimise attack surfaceID.AM Asset Management; PR.AA Identity Management, Authentication, and Access ControlCM-7 Least Functionality; AC-6 Least Privilege
9 Continuous assuranceID.IM Improvement; DE.CM Continuous MonitoringCA-2 Control Assessments; CA-7 Continuous Monitoring; RA-5 Vulnerability Monitoring and Scanning
10 Change securelyPR.PS Platform SecurityCM-3 Configuration Change Control; CM-4 Impact Analyses

For the general method of reusing one control across several frameworks, see the Control Mapping Guide. For threat modelling technique behind principle 6, see the Threat Modelling Guide.

Common pitfalls

  • Filling the tracker at the end: completing every phase sheet just before a beta assessment defeats the purpose. The guidance expects updates as part of regular project routines.
  • Treating a high confidence profile as assurance: the tracker itself says a high profile does not guarantee security and it does not replace existing assurance or the risk register.
  • Accepting supplier claims unchecked: the third-party activity asks whether certificates and assessments are valid and done by impartial assessors, and expects a living risk report that includes rejected products. Skipping this leaves principle 2 unevidenced.
  • Logging humans only: services pass principle 5 on paper while integrations and deployment identities act unlogged. Machine activity needs the same telemetry and alerting.
  • Forgetting retirement: components are decommissioned but their accounts, keys and DNS records remain. The "Retiring service components securely" activity exists for this reason.
  • One controls set per project: every team picking its own frameworks multiplies effort and evidence formats. Agree an organisational set and taxonomy once.

Practitioner checklist

  • Confirm whether your organisation is in Group 1 or Group 2 and that its implementation schedule is agreed with the Government Cyber Unit.
  • Name the CDIO as accountable for adoption and an SRO and service owner for each in-scope service.
  • Complete the RACI for security tasks, with one accountable person per task.
  • Agree an organisational security controls set and map it to CAF outcomes.
  • Start the Self Assessment Tracker in Discovery and update it as evidence changes, not at gates.
  • Run third-party product reviews for every platform, library and SaaS tool, and keep a living risk report.
  • Inventory every non-human identity the service creates, with owner, scope and credential type.
  • Test logging, monitoring and alerting, including machine identity activity, before Public beta.
  • Route every significant change through a security impact evaluation and update the tracker.
  • Store the tracker and its evidence as restricted assets.

Standards and references

Related NHI Mgmt Group resources: NCSC Cyber Assessment Framework (CAF) Guide · Software Liability and Secure by Design Policy Guide · Threat Modelling Guide · Non-Human Identity (NHI) Guide · Control Mapping Guide. This guide summarises UK government Secure by Design policy and guidance and is not legal advice.

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 27 October 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.