CVE-2026-0770: Untrusted Component Execution in Langflow Actively Exploited in the Wild

What the Vulnerability Is

CVE-2026-0770 is an Inclusion of Functionality from Untrusted Control Sphere flaw (CWE-829) in Langflow. It allows unauthenticated or low-privileged actors to trigger execution of code, components, or custom nodes that originate from an untrusted control sphere. In practice, this means an attacker can cause the Langflow server to load and execute logic it should never trust — turning a flow-building feature into a remote code execution primitive.

Affected Software

• Vendor/Product: Langflow
• Affected versions: Not specified in available data — check vendor advisories for exact version ranges and confirm patch status before assuming exposure.

Exploitation Status

CVE-2026-0770 is listed in the CISA Known Exploited Vulnerabilities (KEV) catalog, has a public exploit-db proof-of-concept, and is being actively exploited in the wild. Disclosed 2026-05-29, with no patch date currently recorded. For defenders, this combination — KEV listing, public PoC, and active exploitation — means this should be treated as an immediate priority: patch or mitigate now, and assume opportunistic scanning and exploitation attempts are already occurring against internet-facing Langflow instances.

How Our Detection Catches It

We ship coverage for this technique across Microsoft Sentinel (KQL), Splunk (SPL), Elastic (EQL), IBM QRadar (AQL), Sumo Logic, Google Chronicle (YARA-L), and CrowdStrike (CQL). At a high level, the detection logic focuses on the observable exploitation chain rather than a single indicator:

• Unexpected outbound connections initiated by the Langflow host to fetch remote component or flow definitions.
• Anomalous child process spawns originating from the Langflow server process — a strong signal that untrusted code has been executed.
• Webhook or API calls to Langflow's flow-execution and custom component endpoints arriving from untrusted or unexpected source addresses.

Correlating these signals — network egress, process lineage, and API access patterns — lets the detection distinguish legitimate Langflow component usage from exploitation of CVE-2026-0770, reducing false positives while catching both scripted PoC usage and more tailored attacker tradecraft.

Learn More

For the full detection logic, including ready-to-deploy KQL, SPL, and other SIEM queries, see the CVE-2026-0770 detection page.