Zero Radar
← feed
CriticalVulnerabilityCVE-2026-107328updated 29 Oct 2026, 20:04 UTC

IBM Langflow: logged-in users can run arbitrary code on the server

Langflow is a tool for building AI workflows. A flaw lets someone with a login to the software run their own commands on the server it runs on, which could take over the machine.

Official, 2 sources
SeverityCriticalCVSS 9.9 of 10
Being exploited?UnclearNo reports either way yet
Fix available?UnclearNo fixed version published in the sources yet
How sure is this?Official, 2 sourcesBacked by an official source

How the attack works simplified

  1. 1Attacker logs inNeeds any valid Langflow account
  2. 2Malicious input sentCrafted code reaches the app
  3. 3Code passed to eval/execInput is not sanitized before running
  4. 4Commands run on serverRuns with the app's privileges
  5. ResultAttacker controls the server

What happened

IBM Langflow OSS versions 1.0.0 through 1.12.4 fail to neutralize code passed into eval/exec, allowing remote authenticated code execution.

Why it is dangerous: An attacker with any valid account could gain control of the server, reading or changing data and reaching other systems it can access. Organizations running self-hosted Langflow with shared or self-registered accounts are most at risk.

Are you affected?

You are affected if you run IBM Langflow OSS version 1.0.0 through 1.12.4 and any user can log in to it.

IBM Langflow OSS 1.0.0 through 1.12.4

What to do

  1. Check whether you run Langflow OSS between 1.0.0 and 1.12.4
  2. Restrict who can create or log in to Langflow accounts until patched
  3. Apply the fixed version from IBM or the GitHub advisory as soon as it is available
  4. Review server logs for unexpected commands or new processes

Timeline

  1. 29 Oct 2026, 16:17 UTC
    Publicly disclosed
  2. 29 Oct 2026, 16:17 UTC
    Reported by NVD · NVD
  3. 29 Oct 2026, 16:20 UTC
    Picked up by Zero Radar
  4. 29 Oct 2026, 20:04 UTC
    GitHub Advisory reported on it · GitHub Advisory

Sources 2 reports, oldest first

Official vendor, government CERT or advisory database

Technical details

ID
CVE-2026-107328 · GHSA-5c8g-pvxp-6cx4
CVSS
9.9 / 10
EPSS
0.9% chance of exploitation within 30 days
Disclosed
29 Oct 2026, 16:17 UTC
On Zero Radar
29 Oct 2026, 16:20 UTC (3 min after disclosure)
Last change
29 Oct 2026, 20:04 UTC

The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works

IBM Langflow: logged-in users can run arbitrary code on the server · Zero Radar