CriticalVulnerabilityCVE-2026-107329updated 29 Oct 2026, 20:04 UTC
IBM Langflow allows remote code execution by logged-in users
A flaw in IBM Langflow, a tool for building AI workflows, lets someone who has a login run their own commands on the server. That can mean taking control of the machine.
Official, 2 sources
| Severity | CriticalCVSS 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 |
What happened
IBM Langflow OSS versions 1.0.0 through 1.12.4 have improper control of code generation, allowing a remote authenticated attacker to execute arbitrary code.
Why it is dangerous: An attacker with any valid account could run arbitrary code on the server, potentially reading data, changing systems, or moving further into the network. Organizations exposing Langflow to users or the internet are most at risk.
Are you affected?
You are affected if you run IBM Langflow OSS version 1.0.0 through 1.12.4.
IBM Langflow OSS 1.0.0 through 1.12.4
What to do
- Check whether your Langflow OSS version is between 1.0.0 and 1.12.4
- Restrict who can create or log into Langflow accounts until a fix is applied
- Follow IBM's security advisory for the fixed version and upgrade as soon as it is available
Timeline
- 29 Oct 2026, 16:17 UTCPublicly disclosed
- 29 Oct 2026, 16:17 UTCReported by NVD · NVD
- 29 Oct 2026, 16:20 UTCPicked up by Zero Radar
- 29 Oct 2026, 20:04 UTCGitHub Advisory reported on it · GitHub Advisory
Sources 2 reports, oldest first
Official vendor, government CERT or advisory database
- CVE-2026-107329NVD · 29 Oct 2026, 16:17 UTC
- IBM Langflow OSS 1.0.0 through 1.12.4 could allow a remote authenticated attacker to execute...GitHub Advisory · 29 Oct 2026, 18:30 UTC
Technical details
- ID
- CVE-2026-107329 · GHSA-34r7-8gr9-84jp
- 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