Langflow Smart Transform lets a crafted prompt run code on the server
- Product
- Langflow
- CVE
- CVE-2026-7700
- Affected
- langflow 1.3.0 to 1.10.2
- Fixed in
- 1.10.3
Last verified against sources: October 26, 2026
Am I affected?
Langflow: affected langflow 1.3.0 to 1.10.2. Fixed in 1.10.3.
- Check your version: run
pip show langflow. Versions >= 1.3.0, < 1.10.3 of langflow are affected. [1]
What to do now
- Upgrade to the fixed version: run
pip install --upgrade "langflow>=1.10.3"(first patched version 1.10.3). [1] - Upgrade to Langflow 1.10.3 or later. [1]
- Until you can upgrade, remove Smart Transform from runnable flows. [1]
- Restrict flow creation, editing and execution to trusted users. [1]
- Do not route untrusted or externally controlled data through Smart Transform. [1]
General practice, not from the source: if the affected software could read tokens or credentials, rotate them after upgrading; pin versions and review changes before updating tools your agents run.
What happened
- Langflow versions 1.3.0 through 1.10.2 have a code-injection flaw in the Smart Transform component (LambdaFilterComponent). [1]
- Smart Transform asks a language model to write a one-line Python lambda, then runs the model's answer inside the Langflow process. [1]
- The only check is that the text starts with "lambda" and contains a colon, and the code is then run with eval(). [1]
- A malicious flow author can trigger this directly through the Instructions field. [1]
Impact
- An attacker can make the model return code such as lambda x: __import__("os").system("id"), which passes the format check and runs a command. [1]
- The code runs with the privileges of the Langflow service process. [1]
- This can expose or change credentials, files, application data and network resources the process can reach. [1]
- In shared deployments it may also affect other tenants. [1]
What was done
Sources
- GHSA-9fpm-3445-2vx4 — GitHub Advisory Database
Every statement above was checked against these sources before publishing. Spotted an error? Tell us and we will correct it with a dated note.
Share advisory brief with your team
Copy a plain-text summary for Slack, GitHub issues, or team security channels.
[CVE-2026-7700] Langflow Smart Transform lets a crafted prompt run code on the server
Product: Langflow | Affected: langflow 1.3.0 to 1.10.2 | Fixed in: 1.10.3
Summary: Langflow versions 1.3.0 through 1.10.2 have a code-injection flaw in the Smart Transform component (LambdaFilterComponent).
Source & remediation on AgentArmorHQ: https://agentarmorhq.com/incidents/2026/10/cve-2026-7700-langflowCorrections
Report an error or request a correction. We review requests within 5 business days and add a dated correction note.
Discussions for CVE-2026-7700
Ask questions or share configuration tips. Governed by our Community Rules & Terms.
No posts match this filter yet.
Be the first to ask a question or share a working tip.