2026 brought at least three maximum-severity, unauthenticated RCE vulnerabilities to Magento and Adobe Commerce. Here's the full timeline, what's actually new about this pattern, and the defences that matter more than patch speed alone.

If you've managed a Magento or Adobe Commerce store at any point in 2026, you'll know this year has been relentless. By our count, the platform saw at least three maximum-severity, unauthenticated remote-code-execution vulnerabilities (PolyShell, StyleSmuggler and BlockStrap) on top of four separate monthly and emergency security bulletins fixing more than 30 additional CVEs between them. We've covered each one individually as it happened. This post is the step back: the full timeline in one place, what's genuinely different about this pattern compared with previous years, and the defences that actually matter when patch speed alone isn't enough.
| Date | Event | CVSS | Notes |
|---|---|---|---|
| 12 May | Magento 2.4.9 goes GA | N/A | Ships the PolyShell fix for new installs only |
| 14 July | PolyShell backport, APSB26-73 | N/A | CVE-2026-48356: fix finally reaches 2.4.7/2.4.8, two months after 2.4.9 shipped it |
| 11 August | APSB26-92 | 9.1 (lead CVE) | Seven CVEs, five Critical; lands the same day Magento 2.4.6 loses standard support |
| 4 September | StyleSmuggler exploitation begins | 10.0 | CVE-2026-75650: a true zero-day, exploited in the wild before any patch existed |
| 7 September | Emergency hotfix, APSB26-146 | N/A | Adobe's out-of-band fix for StyleSmuggler, three days after exploitation began |
| 8 September | APSB26-138 | up to 9.3 | Eight CVEs, all Critical, six needing no authentication, one day after the emergency hotfix |
| 13 October | APSB26-153 | up to 10.0 | Eleven CVEs, including the fix for what would shortly be named BlockStrap |
| 15 October | BlockStrap research published | 10.0 | CVE-2026-103849: Sansec names and details it two days after Adobe had already patched it |
| 27 October | PolyShell added to CISA's KEV catalogue | N/A | Confirms active exploitation of a flaw that had a fix available since July |
We've been doing this long enough to remember SessionReaper (CVE-2025-54236) last September, and the comparison is instructive. SessionReaper was a single, very serious incident: a patch leak that accelerated active exploitation of one critical flaw. What 2026 has shown is a pattern: three separate maximum-severity, zero-authentication RCE vulnerabilities in five months, two of them (StyleSmuggler and BlockStrap) rooted in a structurally similar weakness in how Magento's template engine handles content it's told is safe at one rendering stage and re-evaluates at another.
The other genuine shift is speed. Attackers are now reverse-engineering patches and building working exploits within hours to days of a fix becoming public. Sansec's own research noted working BlockStrap exploits reproduced within 24 hours of the October patch. "We'll patch on our usual monthly cycle" was a defensible position a few years ago. In 2026, it isn't.
Honestly, both play a role, and it's worth being straight about it rather than being defensive. Magento and Adobe Commerce run a meaningful share of mid-market and enterprise ecommerce, which makes them a genuinely attractive target: attackers go where the installed base and the payment data are. At the same time, a codebase with Magento's age and extensibility (custom modules, third-party extensions, years of accumulated template logic) has more surface area than a newer, more constrained platform. Sansec's research has become the de facto early-warning system for this entire ecosystem in 2026 precisely because there's enough happening to need one.
Patch speed matters, but it can't be your only line of defence against a true zero-day: by definition, nothing patches you before exploitation begins. In rough order of effort to implement:
No. We want to be unambiguous about that, because it's a dangerous assumption to let slide. A WAF rule blocks a known attack pattern; it doesn't fix the underlying flaw, and a sufficiently different exploit variant can sometimes slip past it. Every vendor whose rules we've mentioned this year (Sansec, Cloudflare, Imperva, Adobe's own Fastly setup) frames their protection as a bridge to patching, not a substitute for it, and that's exactly how we'd recommend you treat it too.
It's been a genuinely demanding year, and we won't pretend otherwise. What we can say honestly is that Deploy has been in the middle of every one of these advisories as they happened: not reading about them after the fact, but testing fixes against our own clients' exact configurations the same day they landed, several times within hours of an emergency hotfix going live. We've done that alongside a number of other leading Magento agencies who've taken this year just as seriously, and we're genuinely proud that the work (including specific contributions from Scott Robinson, our technical lead) has been recognised more than once in 2026, both by Sansec directly and within the wider Magento agency community. That recognition matters to us less as a pat on the back and more as confirmation that the unglamorous, out-of-hours patching discipline this year has demanded is actually the right way to run a Magento support practice.
If you want a second opinion on whether your own store (or your whole portfolio, if you're an agency) is actually covered against everything in the timeline above, our support team can run through it with you. And if 2026 has you questioning whether Magento is still the right platform for your business, that's a different, longer conversation worth having properly rather than deciding on the back of a hard year; we'd rather talk it through than have you make that call based on headlines alone.
Sources: Sansec's StyleSmuggler research, Sansec's PolyShell research, Sansec's BlockStrap research, and Adobe's official security bulletins linked throughout each post above.
Is Magento less secure than other platforms, or just more targeted? Both factors are real: a large installed base makes it an attractive target, and two decades of extensibility give it more surface area than newer, more constrained platforms. Covered in more detail above.
What's the difference between a CVE being "patched" and being "fixed in the wild"? A patch being available and a store actually running it are two different things: PolyShell had a fix in Magento 2.4.9 from May, but most stores on 2.4.7/2.4.8 weren't actually protected until the July backport. "Patched" should mean "verified applied to your specific store," not "a fix exists somewhere."
Does a WAF replace the need to patch? No. It's a bridge that buys you time between disclosure and a tested deploy, not a permanent substitute. See the section above for why.
This page will be updated as new CVEs land through the rest of 2026 and into 2027, so check back, or follow our blog, for the latest entry in the timeline.