What happened
Sansec recorded the first confirmed StyleSmuggler exploitation at 22:20 UTC on September 4, 2026. On September 5, 2026, Sansec disclosed the issue after reproducing the chain on clean Magento Open Source 2.4.7, 2.4.8 and 2.4.9 installations. Sansec reported that the first affected store ran Magento Open Source 2.4.6-p15 with the July and August 2026 patches applied and security:patch-status clean.
The attack path uses POST requests to /graphql with styles parameters and triggers Magento’s Payment Transaction Failed Reminder rendering path. The two-stage chain first poisons a Magento-managed file with PHP and then causes the platform to include that file while rendering the failed-payment message. Opening or successfully delivering the email is not required for execution.
The published implant SHA-256 is b79dfdc1eed860e0b76c629d6adfce251db379b0b45a6d728d4ef483f7551420. Sansec published 99.84.67.186 as the backdoor command-and-control address. Published artefacts include ~/.local/share/.gvfsd/gvfsd-user, /tmp/.kw_
At the September 7, 2026 06:00 America/New_York publication cut-off, Adobe had not posted a StyleSmuggler advisory, CVE identifier or official patch on its Commerce security bulletin index. Liquid Web had activated mitigations, scanning, quarantine and restoration work for affected environments. Attribution posture: Sansec and the cited provider reporting named no threat actor responsible for StyleSmuggler exploitation.
Why this matters now
The decisive weekend change is not another critical score; it is confirmed exploitation without a vendor remediation path. Sansec reported that a store showing a clean patch status was compromised and reproduced the chain across current Magento releases. Normal vulnerability-management evidence therefore cannot establish safety.
Retailers with headless, PWA or integration-heavy storefronts face a business trade-off because GraphQL may be operationally important. Disabling or filtering it can reduce exposure but may interrupt revenue-generating services. That exception requires an explicit owner, time limit, monitoring plan and rollback decision rather than an informal technical workaround.
The backdoor provides persistent command execution after initial exploitation. Applying a later Adobe patch, when available, will not remove an implant already installed. Incident response and application operations must therefore separate entry-point mitigation from host-level compromise assessment.
The decision for security leaders
Assign application security and commerce engineering to choose containment based on storefront architecture. Where GraphQL cannot be disabled, document the compensating control, monitoring owner, expiry and accepted business exposure.
Assign incident response to perform host-level compromise assessment rather than relying on Magento patch-status output. Any published implant, process, cron entry or command-and-control connection should move the system into containment and rebuild procedures.
Require executive approval before restoring a compromised store. Restoration evidence should cover clean infrastructure, rotated secrets, reviewed integrations, validated code and an explicit plan to replace interim controls with Adobe-supported remediation.
Evidence of closure
- Asset register identifies every internet-facing Magento and Adobe Commerce instance.
- Host review records a disposition for every published implant artefact.
- Network evidence shows no connection to 99.84.67.186.
- Approved architecture record documents the interim GraphQL control and expiry date.
The Security.io assessment
The evidence supports confirmed exploitation and a reproducible unauthenticated code-execution chain, but Adobe had not established an official affected-version boundary or remediation at the cut-off. Sansec’s affected-version conclusion should therefore drive urgent precautionary action while remaining clearly attributed to the research source.
The cited sources did not publish a CVE identifier or an Adobe-supported fixed version by the publication cut-off. Unofficial patches may reduce specific paths, but they should not be represented to leadership as vendor fixes or as proof that an already exposed host is clean.
This story warrants inclusion because the weekend development removed the ordinary patch-management response: fully updated stores were reported compromised and the available choices involved service-impacting containment or continued exposure. That is an immediate business-risk decision, not simply an application-security queue item.
Questions for the morning meeting
- Which internet-facing stores run Magento Open Source or Adobe Commerce?
- Can the business disable GraphQL without taking headless or PWA storefronts offline?
- Has every store been checked for the published implant and persistence artefacts?
- Who accepts the residual risk of operating without an Adobe-supported fix?