Critical Java application security | Updated 28 July 2026

CVE-2026-16723 is a critical remote-code-execution vulnerability in Fastjson 1.x. Exposure depends on a specific combination of library version, configuration and Spring Boot executable fat-jar deployment. That makes disciplined inventory more useful than panic.

This Australian action briefing separates verified facts from assumptions, shows exactly what to check and gives technology leaders a practical response sequence.

● Evidence checked against primary sources on 28 July 2026
Australian business leaders reviewing identity and access security controls in a modern Sydney office
A practical identity and access-security review for an Australian business. Compuloop editorial image.

Fastjson RCE at a glance

Confirmed: Fastjson 1.2.68–1.2.83 is affected under the conditions described by the maintainer. The Alibaba CNA supplied a CVSS 3.1 score of 9.0 to the NVD record.

Important qualifier: the issue requires a Spring Boot executable fat-jar deployment with SafeMode off. A version match alone does not prove that an application is exploitable.

Not confirmed: we found no primary-source evidence of a successful Australian compromise. CVE-2026-16723 was not in CISA’s Known Exploited Vulnerabilities catalogue when checked on 28 July 2026.

9.0Critical CVSS 3.1 score supplied by the Alibaba CNA in the NVD record.
1.2.68–1.2.83The Fastjson 1.x versions identified as affected by the maintainer.
84,700+Cybercrime reports received by ASD’s ACSC in FY2024–25 — one every six minutes.
+28%Increase in publicly reported vulnerabilities during FY2024–25, according to ASD’s ACSC.

Australian figures are from the ASD Annual Cyber Threat Report 2024–25. They describe the wider risk environment, not incidents caused by this Fastjson vulnerability.

What exactly is affected?

The Fastjson maintainer advisory defines a narrow but important exposure path. The application must use an affected Fastjson 1.x release, run with SafeMode off and be packaged as a Spring Boot executable fat-jar. The advisory reports verification across Spring Boot 2.x through 4.x and JDK 8, 11, 17 and 21.

Do not use a package version alone as the incident trigger. Confirm the deployment conditions, then prioritise exposed services by internet reachability, privileges, data access and business impact.

CheckAffected conditionPractical decision
Library versionfastjson 1.2.68 through 1.2.83Confirm direct and transitive dependencies with build files, SBOMs and container scans.
ConfigurationAutoType off and SafeMode off — the stock defaults described by the maintainerDo not assume the default configuration is protective for this issue.
Deployment modelSpring Boot executable fat-jarVerify how the application is packaged and launched; version alone does not establish exposure.
Not affected by this CVEAll fastjson2 versions; SafeMode enabled; noneautotype builds; deployment models outside the advisory’s trigger conditionsRecord the evidence and continue normal vulnerability management rather than applying an emergency change blindly.
GitHub advisory header for Fastjson advisory
Official Fastjson advisory asset hosted on GitHub — primary technical source for the vulnerability description

Why Australian business leaders should care

Fastjson may sit several layers below the application a business owner recognises. It can arrive through a direct dependency, a third-party component, an older internal service or a container image. Without a usable software inventory, a technically simple question — “Do we run this?” — can consume valuable response time.

The ASD Annual Cyber Threat Report 2024–25 says publicly reported vulnerabilities rose 28%. It also reports an average self-reported business cybercrime cost of $80,850 per report. Those figures are not Fastjson loss estimates; they show why inventory, patching and evidence-led triage deserve executive attention before the next urgent advisory lands.

Australian cybercrime cost chart showing average losses for small medium and large businesses in 2024 to 2025
Average self-reported cost per business cybercrime report in FY2024–25. Source: ASD Annual Cyber Threat Report 2024–25.

Prioritise the business impact, not just the CVE

  • Internet-facing services: start with APIs and applications that accept untrusted JSON from customers, partners or public endpoints.
  • Identity and privilege: elevate services that can access authentication, secrets, customer records or administrative systems.
  • Operational dependency: identify applications that support revenue, logistics, customer service or regulated workloads.
  • Third-party ownership: obtain written evidence from software vendors when the affected library is embedded in a supplied product.
  • Detection coverage: prioritise systems with weak process, network or application logging because compromise may be harder to rule out.
  • Recovery readiness: confirm clean backups, rebuild instructions and credential-rotation plans before making disruptive changes.
Australian cyber security specialists reviewing incident-response evidence and containment actions
Security specialists reviewing incident evidence and the next containment actions. Compuloop editorial image.

A practical response plan

PriorityActionOwner and evidence
P0 — todayFind fastjson 1.x, identify internet-facing JSON endpoints and confirm fat-jar deployment conditions.Application owner plus security. Preserve dependency scan and deployment evidence.
P0 — containEnable SafeMode or move to the maintainer’s noneautotype build when immediate migration is not practical.Engineering. Test the exact production workload and retain rollback steps.
P1 — remediatePlan and test migration to fastjson2, which the maintainer says is not affected by this vulnerability.Engineering and change management. Validate behaviour in staging before production.
P1 — huntReview Java child processes, unusual egress, new files, scheduled tasks and suspicious JSON containing @type.Security operations. Preserve logs and host evidence before disruptive clean-up.

Detection and investigation pointers

For applications that meet the affected conditions, review telemetry around the first date of exposure and any suspicious requests. Useful evidence includes web and API request logs, Java process creation, DNS and outbound connections, new files in temporary or application directories, service changes, scheduled tasks and privileged-account activity.

Evidence first: do not wipe or rebuild a potentially affected host before preserving logs, volatile evidence and a timeline. If you find unexpected Java child processes, persistence, credential access or lateral movement, engage incident response early.

Searching for @type in requests can support hunting, but it is not a complete detection strategy and should not be treated as proof by itself.

Control lessons beyond this vulnerability

  • Software inventory and SBOMs: record direct and transitive dependencies for production applications and container images.
  • Patch governance: define who can assess, test, approve and deploy emergency security changes.
  • Least privilege: limit service accounts, filesystem access, secrets and network paths to reduce blast radius.
  • Application logging: capture request, process and egress evidence centrally enough to support a time-critical investigation.
  • Supplier assurance: require vendors to answer dependency and remediation questions with evidence, not generic reassurance.
  • Recovery validation: test that critical applications can be rebuilt cleanly and restored without reintroducing vulnerable components.

Turn this advisory into a repeatable response

Compuloop can help map applications and dependencies, prioritise exposure, validate logging and backup readiness, and build a remediation plan your technical team and leadership can both use.

Review cybersecurity services   |   Review Essential Eight services   |   Review backup and recovery

Need a second set of eyes on Fastjson exposure?

Bring us the application list, build artefacts or vendor response. We will help turn the uncertainty into a prioritised action list.

Request a vulnerability review   |   Call 1300 007 613   |   Email sales@compuloop.com.au

Fastjson CVE-2026-16723 questions

Does every use of Fastjson 1.2.68–1.2.83 mean the application is exploitable?

No. The maintainer describes additional configuration and deployment prerequisites. Confirm SafeMode state and the Spring Boot executable fat-jar condition before concluding exposure.

Is Fastjson2 affected?

The maintainer states that all Fastjson2 versions are not affected by this vulnerability because the root cause was removed architecturally. Normal secure configuration and dependency management still apply.

What is the fastest temporary mitigation?

The maintainer lists enabling SafeMode or switching to the 1.2.83 noneautotype build as P0 actions. Test any change against the real application before production deployment.

Has this CVE been confirmed as exploited in Australia?

We found no primary-source confirmation of a successful Australian compromise when this article was checked on 28 July 2026. Absence of public confirmation is not evidence that an individual environment is safe; assess your own exposure and logs.

Should a business owner handle this personally?

The owner should ensure accountability and business priority. Technical validation belongs with the application owner, security team, software vendor or a qualified specialist.

Primary sources and editorial note

This article is general cybersecurity information for Australian organisations, not a claim that any named organisation has been compromised. Compuloop has no access to private Fastjson investigations. Statistics describe the wider Australian threat environment and are not attributed to this vulnerability. The article will be corrected or updated if primary sources materially change.