highAgent ThreatFramework Vulnerability

CISA KEV Catalog Addition Including Langflow Authorization Bypass

First seen Jul 8, 2026 · Updated Jul 8, 2026

CISAKEVLangflowauthorization-bypassagent-frameworkknown-exploitedAI-pipeline-toolASI02 · Tool MisuseSurface: Tool LayerPropagation: Single Hop

CISA added three actively exploited vulnerabilities to its Known Exploited Vulnerabilities catalog, including CVE-2026-55255, an authorization bypass in Langflow, a popular framework for building AI/LLM agent workflows. This is relevant to AI agent security because Langflow is used to construct agentic pipelines, and an authorization bypass could let attackers manipulate or hijack agent flows, credentials, or connected tools. The other two CVEs (JoomShaper and Joomlack) are unrelated CMS plugin vulnerabilities with no agentic relevance.

Technical Analysis

CVE-2026-55255 is described as an 'Authorization Bypass Through User-Controlled Key' vulnerability in Langflow, a low-code framework for building LLM-powered agent and RAG pipelines. This class of bug typically allows an attacker to supply or manipulate a key/identifier (e.g., API key, session token, or object reference) that the server trusts without proper server-side validation, enabling access to other users' flows, credentials, or backend resources. In an agentic context, this could allow an attacker to view or modify agent workflow definitions, exfiltrate embedded API keys/secrets used by connected tools or LLM providers, or inject malicious logic into a shared pipeline that other users/agents execute. Because Langflow orchestrates calls to external tools, models, and data sources, a bypass at the authorization layer crosses from a simple access-control flaw into a supply-chain-style risk affecting every downstream agent or tool invoked by the compromised flow. The KEV listing confirms active exploitation in the wild, elevating urgency beyond a theoretical framework bug.

Affected Systems

Langflow, JoomShaper SP Page Builder, Joomlack Page Builder

Detection Signatures

  • Unexpected access to Langflow flow IDs/objects not owned by the requesting user or API key
  • API requests to Langflow endpoints with manipulated or guessed key/ID parameters
  • Anomalous retrieval of stored credentials or flow configs via authorization-sensitive endpoints
  • Log entries showing cross-tenant or cross-user flow access in multi-tenant Langflow deployments
  • Unexplained modification of agent/tool pipeline definitions

Remediation Steps

  1. 1

    Patch Langflow immediately

    Upgrade to the fixed Langflow version addressing CVE-2026-55255 per vendor advisory; treat as urgent given confirmed active exploitation and KEV listing.

  2. 2

    Audit authorization logic

    Review Langflow deployment for any user-controlled keys/IDs used in authorization decisions; ensure server-side ownership checks, not client-supplied identifiers, gate access to flows and secrets.

  3. 3

    Rotate exposed credentials

    Rotate any API keys, LLM provider tokens, or connected tool credentials stored within Langflow flows that may have been exposed via the bypass.

  4. 4

    Restrict public exposure

    Per BOD 26-04 guidance, prioritize remediation on publicly exposed Langflow instances that grant significant control post-exploitation; restrict network access where possible.

  5. 5

    Check for prior compromise

    Review logs for signs of exploitation prior to patching, consistent with BOD 26-04 requirements for compromise assessment before/after patch application.

CVE / Advisory IDs

CVE-2026-48908CVE-2026-55255CVE-2026-56290

Industries Most Exposed

Federal governmentsoftware developmenttechnologyany organization using Langflow for AI agent orchestration

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.