Servers and applications
CVE-2026-33873
Authenticated Code Execution in Agentic Assistant Validation
- Severity
- Critical 9.3
- Published
- 24 Mar 2026
- Last changed
- 17 Jun 2026
- Added to UpHeron
- 1 Nov 2026
- Fixed in
- 1.9.0
- Exploit chance
- 1.5% in the next 30 daysHigher than 74% of CVEs. EPSS by FIRST, 19 Nov 2026
- Weakness
- Code injection (CWE-94)Improper input validation (CWE-20)Eval injection (CWE-95)Improper access control (CWE-284)
- Attack
- Over the network · Needs a user account · No user action needed
- Attacker can
- Read data · Change data
- CVSS vector
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X- Also known as
- GHSA-v8hw-mh8c-jxfc
- Sources
- NVD, Project advisories (GitHub)
Description
The Agentic Assistant feature in Langflow executes LLM-generated Python code during its validation phase. Although this phase appears intended to validate generated component code, the implementation reaches dynamic execution sinks and instantiates the generated class server-side.
In deployments where an attacker can access the Agentic Assistant feature and influence the model output, this can result in arbitrary server-side Python execution.
The Agentic Assistant endpoints are designed to help users generate and validate components for a flow. Users can submit requests to the assistant, which returns candidate component code for further processing.
A reasonable security expectation is that validation should treat model output as untrusted text and perform only static or side-effect-free checks.
The externally reachable endpoints are:
The request model accepts attacker-influenceable fields such as input_value, flow_id, provider, model_name, session_id, and max_retries:
Affected software
| Software | Affected versions |
|---|---|
LangflowServers and applications · cpe:a:langflow:langflow |
|
LangflowServers and applications · github:langflow-ai/langflow |
|
About Langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
- Weekly downloads (PyPI)
- 8,374PyPI Stats, 15 Nov 2026
- Docker Hub pulls
- 4,731,179Docker Hub, 15 Nov 2026
- GitHub stars
- 156,656GitHub, 15 Nov 2026
- Forks
- 10,412GitHub, 15 Nov 2026
- Latest release
- v1.13.2GitHub, 15 Nov 2026
- Released
- 10 Nov 2026GitHub
- License
- MITGitHub
- First published on GitHub
- 8 Feb 2023GitHub
Source code Website Docker image Package
Vulnerabilities in Langflow
- In our list
- 228
- Published in the last 12 months
- 221
- Critical, last 12 months
- 60
- High, last 12 months
- 101
- Known exploited
- 6
- CVE-2026-97677Langflow OSS is affected by multiple vulnerabilitiesHigh 8.1 8 Nov 2026
- GHSA-qvqp-mxh4-f27wLangflow URL component: unbounded response size allows denial of service via memory exhaustionMedium 6.5 5 Nov 2026
- CVE-2026-18894Langflow OSS is affected by multiple vulnerabilitiesCritical 9.8 29 Oct 2026
- CVE-2026-77817Langflow OSS is affected by multiple vulnerabilitiesMedium 6.5 29 Oct 2026
- CVE-2026-77819Langflow OSS is affected by multiple vulnerabilitiesMedium 6.5 29 Oct 2026
- CVE-2026-90991Langflow OSS is affected by multiple vulnerabilitiesMedium 6.5 29 Oct 2026
References
- NVD https://nvd.nist.gov/vuln/detail/CVE-2026-33873
- CVE.org https://www.cve.org/CVERecord?id=CVE-2026-33873
- GitHub Advisory Database https://github.com/advisories/GHSA-v8hw-mh8c-jxfc
- github.com https://github.com/langflow-ai/langflow/security/advisories/GHSA-v8hw-mh8c-jxfc