highAgent ThreatFramework Vulnerability

CISA KEV Catalog Addition Including Langflow Untrusted Code Inclusion Vulnerability (CVE-2026-0770)

First seen Jul 22, 2026 · Updated Jul 22, 2026

CISAKEVLangflowagent-frameworkknown-exploited-vulnerabilityBOD-26-04patch-managementASI04 · Agentic Supply ChainSurface: Supply ChainPropagation: Single Hop

CISA added four actively exploited vulnerabilities to its KEV Catalog, one of which (CVE-2026-0770) affects Langflow, a framework used to build AI agent/LLM workflows, involving inclusion of functionality from an untrusted control sphere. The other three CVEs concern DD-WRT and WordPress and are unrelated to AI agent infrastructure. This is a legitimate, actively exploited vulnerability disclosure relevant to organizations running Langflow-based agent pipelines, warranting prompt patching per federal directive BOD 26-04.

Technical Analysis

CVE-2026-0770 is categorized as 'Inclusion of Functionality from Untrusted Control Sphere,' suggesting Langflow can be induced to load or execute code/components/plugins from an untrusted source, which in an agent-orchestration context could allow an attacker to inject malicious flow components, tools, or logic into a running agent pipeline. Since Langflow is used to visually construct LLM/agent workflows (chains, tools, prompts), this class of vulnerability could enable an attacker to hijack tool execution, exfiltrate credentials/secrets embedded in flows, or achieve remote code execution on the host running the Langflow server. The entry point is likely the Langflow web UI/API accepting untrusted flow definitions, custom components, or imports without sufficient sandboxing or validation, crossing the boundary from configuration/data into executable agent logic. CISA confirms active exploitation in the wild, elevating urgency beyond typical CVSS-only risk scoring. The other three CVEs (DD-WRT, WordPress x2) are standard web/network CVEs with no agentic AI relevance.

Affected Systems

Langflow

Detection Signatures

  • Monitor Langflow server logs for unexpected flow imports, custom component uploads, or execution of components not authored internally
  • Alert on outbound network connections from Langflow host to unfamiliar domains following flow load
  • Review Langflow instance exposure to the public internet; flag any unauthenticated or weakly authenticated admin/API endpoints
  • Audit for unexpected process spawning or file writes originating from the Langflow application process
  • Check for anomalous API calls to /api/v1/ endpoints handling flow or component uploads

Remediation Steps

  1. 1

    Patch Langflow immediately

    Upgrade to the fixed Langflow version addressing CVE-2026-0770 as soon as available from the vendor.

  2. 2

    Restrict network exposure

    Ensure Langflow instances are not publicly accessible without authentication; place behind VPN or strict access controls.

  3. 3

    Validate flow/component sources

    Disable or restrict import of flows, custom components, or code from untrusted or external sources; implement allowlisting.

  4. 4

    Follow BOD 26-04 guidance

    FCEB agencies must remediate per CISA KEV timelines and check for prior compromise indicators before/after patching.

  5. 5

    Audit historical activity

    Review logs for evidence of exploitation prior to patch deployment, consistent with BOD 26-04 compromise-check requirements.

CVE / Advisory IDs

CVE-2021-27137CVE-2026-0770CVE-2026-63030CVE-2026-60137

Industries Most Exposed

Federal governmentTechnologySoftware developmentCritical infrastructure

Sources

Respond to this threat

Pro subscribers get a full AI-generated incident-response playbook for this threat — detection, containment, eradication, and recovery steps — plus an unlimited AI Threat Advisor for questions about your environment.