INCIDENT MAP

Browse / Improper Output Handling / CVE-2026-33873

Langflow has Authenticated Code Execution in Agentic Assistant Validation

CVE-2026-33873 · 2026-03-26 critical

Class
Improper Output Handling (filed by a model that read the advisory)
Technique
Code and command execution
Affected
langflow (PyPI)
CVSS
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
EPSS
1.5% chance of exploitation in 30 days

How it works, in short

Generated component code is dynamically executed during the assistant's validation step, so a user who steers model output can run Python on the server. (Our summary, written by a model. The source text is below.)

Details

Description

1. Summary

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.

2. Description

2.1 Intended Functionality

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:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/api/router.py#L252-L297

The request model accepts attacker-influenceable fields such as input_value, flow_id, provider, model_name, session_id, and max_retries:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/api/schemas.py#L20-L31

2.2 Root Cause

In the affected code path, Langflow processes model output through the following chain:

/assist
→ execute_flow_with_validation()
→ execute_flow_file()
→ LLM returns component code
→ extract_component_code()
→ validate_component_code()
→ create_class()
→ generated class is instantiated

The assistant service reaches the validation path here:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/services/assistant_service.py#L58-L79

The code extraction step occurs here:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/helpers/code_extraction.py#L11-L53

The validation entry point is here:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/helpers/validation.py#L27-L47

The issue is that this validation path is not purely static. It ultimately invokes create_class() in lfx.custom.validate, where Python code is dynamically executed via exec(...), including both global-scope preparation and class construction.

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/lfx/src/lfx/custom/validate.py#L241-L272

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/lfx/src/lfx/custom/validate.py#L394-L399

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/lfx/src/lfx/custom/validate.py#L441-L443

As a result, LLM-generated code is treated as executable Python rather than inert data. This means the “validation” step crosses a trust boundary and becomes an execution sink.

The streaming path can also reach this sink when the request is classified into the component-generation branch:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/services/assistant_service.py#L142-L156

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/agentic/services/assistant_service.py#L259-L300

3. Proof of Concept (PoC)

4. Impact

5. Exploitability Notes

This issue is most accurately described as an authenticated or feature-reachable code execution vulnerability, rather than an unconditional unauthenticated remote attack.

Severity depends on deployment model:

The assistant feature depends on an active user context:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/api/utils/core.py#L38

Authentication sources include bearer token, cookie, or API key:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/services/auth/utils.py#L39-L53

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/services/auth/utils.py#L156-L163

Default deployment settings may widen exposure, including AUTO_LOGIN=true and the /api/v1/auto_login endpoint:

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/lfx/src/lfx/services/settings/auth.py#L71-L87

https://github.com/langflow-ai/langflow/blob/f7f4d1e70ba5eecd18162ec96f3571c2cfbcd1fc/src/backend/base/langflow/api/v1/login.py#L96-L135

6. Patch Recommendation

Discovered by: @kexinoh (https://github.com/kexinoh, works at Tencent Zhuque Lab)

References

Source: GitHub Advisory Database, used under CC BY 4.0.