The Document Foundation LibreOffice · Public exploit code

[Single source] LibreOffice patches spreadsheet flaws that run code without macros

· 1 min read

Security researcher Minseong Kim demonstrated remote code execution in LibreOffice chaining CVE-2026-104692 and CVE-2026-104007: a crafted spreadsheet automatically downloads a malicious database file and runs attacker-controlled Java code when Java is installed and enabled, with no macros or security prompts. No in-the-wild exploitation is reported by the source. Fixes are available in versions 26.2.7 and 26.8.2; update now.

Illustration: LibreOffice patches spreadsheet flaws that run code without macros
WhatsApp LinkedIn X
HighVulnerabilityPublic exploit codePatched

In 30 seconds

Alleged

LibreOffice has shipped security fixes for two flaws that, according to security researcher Minseong Kim, let a crafted spreadsheet run attacker-controlled Java code the moment it is opened. No macros and no security prompts are involved, the report says. The issues, tracked as CVE-2026-104692 and CVE-2026-104007, affect LibreOffice 26.2.6 and earlier plus 26.8.1 on Windows, macOS, and Linux, provided Java is installed and enabled. Versions 26.2.7 and 26.8.2 carry the fixes, and no exploitation in the real world has been reported by the source.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

What happened

Observed

The story surfaced on November 18, 2026 in a public post by DarkWebInformer describing a demonstration by security researcher Minseong Kim. The post credits Kim with videos showing the exploit working on Windows, macOS, and Linux, and it frames the finding as a remote code execution chain, meaning an attack that runs software of the attacker's choosing on the victim's machine.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Alleged

According to the report, the chain links two vulnerabilities tracked as CVE-2026-104692 and CVE-2026-104007. The attack starts when the victim opens a specially crafted spreadsheet. The file then automatically downloads a malicious database file, and that file executes Java code controlled by the attacker. On a system where Java is installed and enabled, the whole sequence reportedly needs no extra clicks, no macro approval, and no security warning.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Why it matters

Inferred

If the demonstration holds up, it undercuts the standard advice that spreadsheet files are safe as long as macros stay disabled. Here the dangerous action is simply opening the file, which is about the lowest bar an attack can set, and users get no prompt they could refuse.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Inferred

There is a gate, though. The execution path described depends on Java being installed and enabled, a setup many desktop users do not have. That narrows the exposed population, yet office environments where Java and LibreOffice coexist remain plausible targets.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

What to do

Alleged

Update LibreOffice now. The report states that security fixes ship in versions 26.2.7 and 26.8.2, so anyone running 26.2.6 or earlier, or version 26.8.1, should move to the matching patched release.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Inferred

As an extra layer, disabling or removing the Java integration in LibreOffice cuts off the reported execution path, and treating unexpected spreadsheet files from untrusted sources with suspicion remains good hygiene regardless of patch status.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Confirmed and still uncertain

Observed

What is on the table so far: a public demonstration with videos for Windows, macOS, and Linux, all credited to Minseong Kim; a list of affected versions, 26.2.6 and earlier plus 26.8.1; and patched releases, 26.2.7 and 26.8.2.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Inferred

What is still open: whether anyone has actually been attacked this way, since the source reports no real-world exploitation; whether the technical claims have been independently verified; and which of the two CVEs handles which step of the chain, a breakdown the public post does not provide.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

What people are saying

Observed

DarkWebInformer, the account that spread the details, called the flaws critical and stressed that opening a spreadsheet is enough, with no macros or security warnings required. Minseong Kim, credited as the researcher behind the find, is the source of the demonstration videos across all three operating systems.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Observed

No statement from The Document Foundation, the organization behind LibreOffice, or from any other party appears in the available material, so the discussion so far is a single reporter relaying a single researcher's work.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Evolution

Inferred

The evidence base is a single snapshot, one post published on November 18, 2026, so the picture has not widened yet. Milestones that would move it: an official LibreOffice security advisory confirming the CVEs and fixes, independent reproduction of the chain, and any first sign of exploitation in the wild.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Reach

Observed

1 independent groups, 1 updates, and 2 engagements.

Technical detail

Alleged

The chain works in two steps, according to the report. A crafted spreadsheet triggers one flaw to automatically download a malicious database file, and a second flaw turns that file into running Java code. The result is remote code execution, attacker software running on the target machine, without macros and without any dialog the user could decline.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Inferred

The two weaknesses are cataloged as CVE-2026-104692 and CVE-2026-104007, but the post does not say which flaw drives the download and which drives the execution. The exact division of labor inside the chain is therefore unspecified in the public record so far.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Sources and evidence

Observed

The available evidence is one post on X by DarkWebInformer, published on November 18, 2026. It describes the demonstration, lists the affected and fixed versions, and references videos for Windows, macOS, and Linux, all credited to researcher Minseong Kim.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Inferred

That is a thin base for a story of this weight: one social media account relaying one researcher's work, with no vendor advisory or independent confirmation in hand yet. Treat the mechanics as demonstrated claims rather than settled fact until corroboration arrives.DarkWebInformerx.com · DarkWebInformer · xOpen original sourceSee in source list

Sources

  1. DarkWebInformerx.com · DarkWebInformer · x

Incidents like this, every morning

The ones that need action first, in a 3-minute read.

Free · 1 email a day, in the morning · no spam, unsubscribe in 1 click

Get the next one tomorrow morning

One email a day with what changed, what to fix first and how.

Free · 1 email a day, in the morning · no spam, unsubscribe in 1 click

Share it with your team

Whoever handles security with you should get it too.

LinkedIn X

Your colleague gets a confirmation request. Nothing is sent unless they confirm.

See all news

Read also

Before you go: get this every day

The HIFY Radar daily edition lands every morning with the incidents that need action. Free.

Free · 1 email a day, in the morning · no spam, unsubscribe in 1 click