CVE-2026-48397

The CVE describes a deserialization flaw in Lightroom where opening a malicious file leads to code execution at the current user's full privilege level. The 'user interaction required' framing in the advisory—that the victim must open a malicious file—obscures more than it reveals. For Lightroom's core user population, opening files from untrusted sources isn't an edge case; it's the job. Photographers receive client deliverables, designers ingest assets from stock sites and collaborators, and production artists process content from external pipelines. The precondition 'victim must open a malicious file' describes routine professional behavior, not a meaningful security barrier. The scope change indicator is the critical technical detail here. When file parsing escapes Lightroom's security boundary and executes at full user privilege, it reveals that Lightroom treats its entire ingestion pipeline as a single trust domain. There's no sandbox between 'parse this RAW file' and 'run code in my user context.' That's a structural shortcut—performance and file-format complexity make isolation expensive, so developers implicitly trusted the file-format layer. That trust is now the vulnerability. More concerning is what happens after exploitation succeeds. Lightroom's value proposition is persistence—the catalog serializes your entire library state, including edits, metadata, collections, and history. A malicious file imported today gets written into that catalog and will be deserialized every time the catalog opens, potentially for years. This isn't a one-shot RCE; it's a persistent-state compromise. The exploit nests in the user's workflow and survives backups, migrations, and reinstalls. The EPSS score of 0.00536 reflects general user populations, but Lightroom users aren't general users. The relevant question isn't 'how likely is any Lightroom user to open a malicious file'—it's 'how likely is a file opened by Lightroom to have originated from an untrusted pipeline.' For creative professionals, that probability approaches unity. Stock downloads, client deliverables, and collaborative platforms are all untrusted ingestion points by any reasonable definition, and Lightroom sits at the center of all of them. What you should check: verify whether Adobe's patch addressed the serialization layer architecturally (sandboxing file parsing or replacing unsafe deserialization patterns) or merely closed a specific gadget chain. Without a before-and-after diff, there's no way to distinguish a genuine architectural fix from a surface patch that leaves the underlying pattern intact—and the history of this vulnerability class in Adobe's creative suite suggests surface patches are the norm. If Lightroom users share catalog files, synchronization presets, or published export collections with collaborators, the blast radius extends beyond the initial victim to anyone consuming that shared workflow state.

Reviewed through automated stages and approved by a human before publication.