A security questionnaire lands from an enterprise prospect. Question 14 asks for your Essential Eight maturity level, and the answer your IT provider gave covers laptops, Microsoft 365 and Office macros. Nobody has looked at the GitHub organisation, the CI runners or the cloud account the pipeline deploys into. That gap is what Essential Eight software development is really about.
It matters because the pipeline is one of the most privileged systems you run. It holds deployment credentials, pulls in third-party code and pushes to production several times a week. A strong endpoint score sitting on top of a weak pipeline gives the board a false sense of safety.
Below we translate the eight strategies into pipeline controls a CTO can assign to someone, measure and evidence.
Note: This article is general information, not legal advice. Check the official guidance and get advice for your situation.
ASD has signalled a change, so why bother now?
The Australian Signals Directorate (ASD) publishes the Essential Eight as a prioritised set of mitigation strategies with a maturity model. Information Age reported on 25 June 2026 that the controls were introduced in 2017 and updated in 2023, and that ASD had released a proposed replacement called "Essentials" for comment. In that proposal the current Essential Eight becomes the first chapter, "Essentials for Enterprise IT", with later chapters progressively covering areas such as operational technology.
On timing, iTnews reported on 24 June 2026 that an ASD official expected a transition period with both documents live, then deprecation of the Essential Eight in probably about 12 months and full retirement in 24 months. The same official said investment made under the Essential Eight would still be relevant under Essentials.
So don't wait. The Essential Eight is still the published baseline, the proposed direction is broader rather than weaker, and every control below carries over to a threat-informed framework.
Essential Eight software development: mapping the eight to your pipeline
The eight strategies are patch applications, patch operating systems, multi-factor authentication, restrict administrative privileges, application control, restrict Microsoft Office macros, user application hardening and regular backups. Some map straight onto engineering work. Others only loosely.
| Essential Eight strategy | What it means in a delivery pipeline | Evidence you can show |
|---|---|---|
| Patch applications | Keep libraries, frameworks and runtimes current; scan dependencies on every build | Scan reports, merged update pull requests |
| Patch operating systems | Rebuild base images and self-hosted runners on a schedule, not only when something breaks | Image build dates, runner patch history |
| Multi-factor authentication | Enforce MFA on source control, CI/CD, cloud consoles and package registries (npm, PyPI, container registry) | Identity provider policy, org-level MFA setting |
| Restrict administrative privileges | Least-privilege pipeline identities, no standing production admin, short-lived credentials | Role definitions, break-glass logs |
| Application control | Only signed artefacts from your own registry can run in production | Admission or deployment policy, signing records |
| Restrict Microsoft Office macros | Mostly an endpoint control; apply the same thinking to unreviewed scripts and automation | Endpoint policy, review on scripts |
| User application hardening | Harden developer browsers and control IDE extensions | Managed device baseline, extension allow list |
| Regular backups | Back up repos, pipeline definitions, IaC state, secrets and data, and test restores | Restore test records |
Treat the table as a starting point, not a compliance claim. An assessor will work from the published maturity model, so check the exact requirements for the level you're targeting.
Patching starts in the build, not the server room
Most application vulnerabilities now arrive through dependencies. A typical Node or Python service has hundreds of transitive packages in its lock file, and any one of them can be the next urgent advisory. If "patch applications" is filed as an ops task, most of the risk sits outside it.
Three habits cover most of it:
- Run dependency scanning on every pull request and on a nightly schedule. The nightly run matters: new advisories land against code that hasn't changed.
- Turn on automated update pull requests (Dependabot or Renovate) and agree a merge window for security updates, for example 48 hours for critical and two weeks for high.
- Rebuild and redeploy container images regularly. A patched
node:22-slimbase image does nothing until you rebuild from it.
The dependency side gets a fuller treatment in our guide to software supply chain security and SBOMs.
Your CI system is an administrator
A pipeline that can deploy to production is, in effect, an admin account. A pattern we see often: one long-lived cloud access key stored as a repository secret, with rights to every environment and no expiry. If it leaks through a build log, a pull request from a fork or a compromised third-party action, the attacker gets all of it.
The fix is well understood. Use OIDC federation so each run gets short-lived credentials (GitHub Actions to AWS via id-token: write and an IAM role trust policy, or Azure workload identity federation). Scope each pipeline identity to one environment. Require MFA and an approval step for any human who can touch production. Review these identities as carefully as staff admin accounts, because that's what they are.
Application control for software teams
On endpoints, application control means only approved programs run. The pipeline equivalent: only artefacts your trusted pipeline built, from reviewed code, can be deployed. In practice that's signing images at build time (Sigstore cosign is the common choice) and having the platform reject anything unsigned or pulled from an unknown registry, for example with a Kubernetes admission policy.
It also kills a shortcut we still see: someone builds an image on a laptop and pushes it straight to production "just this once". With a deployment policy in place, that push fails.
Backups that cover what engineers depend on
Backup scope is often just the production database. A real recovery also needs source repositories, pipeline configuration, Terraform state, secrets and container registries. Lose one of those and a one-day outage can stretch into weeks of rebuilding.
What counts is a tested restore. Schedule one, time it and write down what was missing. Then check whether the credentials an attacker would gain from your pipeline can also delete the backups. If they can, that's the first thing to fix, and it becomes very real during a ransomware incident, as we cover in our piece on ransomware payment reporting.
Going past the baseline: a secure SDLC Australia teams can evidence
The Essential Eight is a set of baseline mitigations, not a full secure development lifecycle. Teams that want a stronger position add design-time and release-time practices:
- Threat modelling for new features and integrations, using something lightweight like our one-afternoon threat modelling workshop.
- Static analysis and secret scanning on every pull request, with push protection turned on so secrets are blocked before they land.
- Security checks for mobile releases, such as those in our mobile app security checklist.
- A short release gate that confirms scans passed, approvals exist and the artefact is signed.
None of this needs a big security team. It needs named owners and controls that run on every build rather than relying on someone remembering.
How JRC Software helps
Our secure development and DevSecOps team starts with a posture review. In the first week we read your CI/CD configuration, list every pipeline identity and what it can touch, check MFA and branch protection on your source control org, and look at how images are built, stored and deployed. We also ask when you last restored from backup, and what went wrong.
You get the table above filled in for your environment, the gaps ranked by risk, and an ordered fix list with effort estimates. We'll also tell you what not to do yet: there's no point buying a signing platform before your pipeline credentials are short-lived. If you want help with the fixes, we work alongside your engineers and leave evidence your security lead can reuse in assessments and questionnaires.
To see where your pipeline actually stands, book a security posture review and we'll walk through it with your team.
Frequently asked questions
Does the Essential Eight apply to software development teams?
Yes, in part. It was written mainly for internet-connected enterprise IT, but patching applications, multi-factor authentication, restricting admin privileges, application control and backups all apply directly to engineering. Source control, CI/CD, cloud accounts and package registries are part of the environment an attacker would target, so include them when you assess maturity.
Is ASD replacing the Essential Eight?
ASD has proposed it. In June 2026 it released a draft called Essentials for comment, and iTnews reported an ASD official saying the Essential Eight would probably start to be deprecated in about 12 months and be retired in 24 months, with both live during a transition. Until final guidance is published, the Essential Eight remains the baseline to work towards.
What is the first pipeline control to fix?
Privileged access, for most teams. Replace long-lived deployment keys with short-lived federated credentials, limit each pipeline identity to one environment, and enforce MFA on source control and CI/CD. That cuts the damage from a leaked secret more than any other single change, and it's usually achievable in a few weeks.
How does the Essential Eight relate to a secure SDLC?
The Essential Eight is a baseline of environment mitigations, while a secure software development lifecycle covers how you design, build, test and release software. Use the Essential Eight for environment controls, then add threat modelling, static analysis, secret scanning, dependency scanning and a release gate to cover the development lifecycle itself.
Cyber Security
Book a security posture review
Tell us where you're stuck and we'll come back with practical next steps.
Get in touch