Skip to content
StratQuad

Secure development practices

2 min read

The UK Software Security Code of Practice, NIST's Secure Software Development Framework and OWASP's verification standards, and how to apply them.

At a glance

Owner
StratQuad
Version
November 2026
Applies to
Organisations that develop, sell or procure software
Mandatory
Voluntary codes and frameworks; increasingly referenced in procurement and in Cyber Essentials

Key points

  • The UK Software Security Code of Practice, published by DSIT in May 2025 and designed with the NCSC, sets out what software vendors should do across 14 principles in four themes.
  • NIST's Secure Software Development Framework (SP 800-218, version 1.1; a draft version 1.2 was published in December 2025) gives a vendor neutral set of practices that most secure development programmes map to.
  • OWASP ASVS turns application security into testable requirements at three levels; OWASP SAMM measures the maturity of the development programme around it.
  • The NCSC Cyber Assessment Framework 4.0 added contributing outcome A4.b, Secure Software Development and Support, and Cyber Essentials v3.3 references the Code of Practice.

The UK Software Security Code of Practice

The Code is voluntary and aimed at software vendors, with a self assessment that buyers can ask for. Its 14 principles sit in four themes:

  • Secure design and development

    Follow an established secure development framework, understand third party components and their risks, test software, and follow secure by design and by default principles.

  • Build environment security

    Protect the build environment against unauthorised access, and control and log changes to it.

  • Secure deployment and maintenance

    Distribute software securely, run a vulnerability disclosure process, detect and fix vulnerabilities promptly, report vulnerabilities to relevant parties, and provide timely security updates.

  • Communication with customers

    Tell customers how long products will receive security updates, give at least a year's notice before support ends, and share information during incidents.

Practices that matter most

PracticeWhat good looks like
Threat modellingDone for new features that change trust boundaries, recorded and revisited
Dependency managementA software bill of materials (SBOM), automated alerts on vulnerable components, a policy for unmaintained packages
Code reviewPeer review on every change, with security checklists for high risk areas such as authentication
Static and dynamic testingSAST in the pipeline, DAST against test environments, findings triaged with owners
Secrets handlingNo secrets in source code; a secrets manager; scanning of repositories and history
Build integrityProtected branches, signed commits or artefacts, isolated build runners, provenance records
Vulnerability disclosureA published policy and security.txt, an acknowledgement time, a remediation process
Support lifecyclePublished end of support dates and security update commitments

Buying software

When procuring, ask suppliers for their Code of Practice self assessment, their SBOM, their vulnerability disclosure policy, and their supported versions and update commitments. These map directly to A4.a and A4.b in CAF 4.0 and to the supplier controls in ISO/IEC 27001 Annex A.

Practitioner note

The cheapest control with the biggest return is dependency management: most exploited application vulnerabilities sit in third party components rather than in an organisation's own code. Start there, then add threat modelling for the few systems that handle the most sensitive data, rather than trying to adopt every practice at once.

Common mistakes

  • Security testing scheduled once before release rather than built into the pipeline.
  • SAST and dependency scanners producing thousands of findings with no triage, so developers ignore them.
  • Secrets removed from code but left in repository history.
  • No published support period, leaving customers unable to plan upgrades.

Sources

Suggest a correction