CVE-2026-41273: Flowise auth bypass exposes OAuth 2.0

CISO Take

Flowise, the popular drag-and-drop LLM workflow builder, contains an authentication bypass (CWE-306) in versions prior to 3.1.0 that lets any unauthenticated attacker retrieve OAuth 2.0 access tokens for every third-party service integrated into your AI workflows. The attack requires zero privileges, no user interaction, is remotely exploitable, and a public proof-of-concept already exists — meaning the exploit bar is at script-kiddie level today. With an EPSS score in the 78th percentile and CISA's SSVC flagging this as TRACK_STAR, this cannot be deprioritized even in the absence of KEV listing. Blast radius extends well beyond Flowise itself: stolen OAuth tokens can grant authenticated access to Google Workspace, GitHub, Slack, databases, or any other service the AI workflow was authorized to use. Patch immediately to Flowise 3.1.0 and rotate all OAuth credentials connected to any pre-patch deployment.

Sources: NVD EPSS GitHub Advisory ATLAS CISA KEV

What is the risk?

HIGH. The CVSS 8.2 (AV:N/AC:L/PR:N/UI:N) score reflects worst-case network accessibility with no barriers to exploitation. Public PoC code removes the skill threshold entirely. The cascading OAuth token theft has a blast radius that extends into every downstream service integrated with Flowise workflows — the initial vulnerability is a 8.2, but the downstream impact on connected enterprise systems could be far higher. Flowise has accumulated 59 prior CVEs, indicating persistent security debt and suggesting this package warrants elevated scrutiny as part of any AI supply chain review. Organizations that have not restricted Flowise configuration endpoints at the network perimeter should treat this as an active threat.

How does the attack unfold?

Reconnaissance

Attacker discovers an internet-exposed Flowise instance via Shodan/Censys scanning and identifies publicly accessible chatflow configuration endpoints.

AML.T0006

Agent Config Extraction

Unauthenticated HTTP GET to the public chatflow endpoint returns the full workflow configuration JSON, including OAuth credential identifiers for every connected integration.

AML.T0084

Credential Theft

Attacker passes extracted OAuth credential identifiers to token refresh endpoints, obtaining valid OAuth 2.0 access tokens for connected third-party services (Google Drive, GitHub, Slack, databases).

AML.T0083

Lateral Movement and Exfiltration

Stolen OAuth tokens provide authenticated access to all services the AI workflow was authorized to use, enabling data exfiltration, unauthorized actions, and persistent re-access via long-lived refresh tokens.

AML.T0091.000

Reconnaissance

Attacker discovers an internet-exposed Flowise instance via Shodan/Censys scanning and identifies publicly accessible chatflow configuration endpoints.

AML.T0006

Agent Config Extraction

Unauthenticated HTTP GET to the public chatflow endpoint returns the full workflow configuration JSON, including OAuth credential identifiers for every connected integration.

AML.T0084

Credential Theft

Attacker passes extracted OAuth credential identifiers to token refresh endpoints, obtaining valid OAuth 2.0 access tokens for connected third-party services (Google Drive, GitHub, Slack, databases).

AML.T0083

Lateral Movement and Exfiltration

Stolen OAuth tokens provide authenticated access to all services the AI workflow was authorized to use, enabling data exfiltration, unauthorized actions, and persistent re-access via long-lived refresh tokens.

AML.T0091.000

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Flowise	npm	—	No patch

Do you use Flowise? You're affected.

How severe is it?

CVSS 3.1

8.2 / 10

EPSS

0.3%

chance of exploitation in 30 days

Higher than 22% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Trivial

Exploitation Confidence

medium

○ CISA SSVC: Public PoC

○ Public PoC indexed (trickest/cve)

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AV Network

AC Low

PR None

UI None

S Unchanged

C High

I Low

A None

What should I do?

6 steps

PATCH

Upgrade Flowise to 3.1.0 immediately — this is the only complete fix.
WORKAROUND (if patching is blocked): Restrict access to chatflow configuration endpoints (/api/v1/public-chatflows/* and related config routes) via WAF rules, reverse proxy ACLs, or firewall at the network perimeter.
ROTATE CREDENTIALS

Assume all OAuth credentials connected to any pre-patch Flowise instance are compromised. Revoke and reissue all OAuth tokens for connected integrations regardless of whether exploitation is confirmed.
AUDIT

Review OAuth grant and access logs in connected platforms (Google Workspace, GitHub, Slack, etc.) for anomalous access in the past 30 days.
INVENTORY

Enumerate all Flowise instances in the environment including developer sandboxes and shadow IT deployments.
DETECT

Alert on unauthenticated GET requests to Flowise chatflow configuration endpoints in web/application logs.

What does CISA's SSVC say?

Decision Track*

Exploitation poc

Automatable Yes

Technical Impact partial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Auth Bypass Data Extraction Agent Framework AML.T0049 - Exploit Public-Facing Application AML.T0083 - Credentials from AI Agent Configuration AML.T0084 - Discover AI Agent Configuration AML.T0091.000 - Application Access Token AML.T0106 - Exploitation for Credential Access

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity

ISO 42001

A.6.1 - Operational and monitoring procedures for AI systems A.7.4 - AI system security

NIST AI RMF

MANAGE 2.2 - Mechanisms to sustain risk management throughout the AI lifecycle

OWASP LLM Top 10

LLM02 - Sensitive Information Disclosure

Frequently Asked Questions

What is CVE-2026-41273?

Flowise, the popular drag-and-drop LLM workflow builder, contains an authentication bypass (CWE-306) in versions prior to 3.1.0 that lets any unauthenticated attacker retrieve OAuth 2.0 access tokens for every third-party service integrated into your AI workflows. The attack requires zero privileges, no user interaction, is remotely exploitable, and a public proof-of-concept already exists — meaning the exploit bar is at script-kiddie level today. With an EPSS score in the 78th percentile and CISA's SSVC flagging this as TRACK_STAR, this cannot be deprioritized even in the absence of KEV listing. Blast radius extends well beyond Flowise itself: stolen OAuth tokens can grant authenticated access to Google Workspace, GitHub, Slack, databases, or any other service the AI workflow was authorized to use. Patch immediately to Flowise 3.1.0 and rotate all OAuth credentials connected to any pre-patch deployment.

Is CVE-2026-41273 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2026-41273, increasing the risk of exploitation.

How to fix CVE-2026-41273?

1. PATCH: Upgrade Flowise to 3.1.0 immediately — this is the only complete fix. 2. WORKAROUND (if patching is blocked): Restrict access to chatflow configuration endpoints (/api/v1/public-chatflows/* and related config routes) via WAF rules, reverse proxy ACLs, or firewall at the network perimeter. 3. ROTATE CREDENTIALS: Assume all OAuth credentials connected to any pre-patch Flowise instance are compromised. Revoke and reissue all OAuth tokens for connected integrations regardless of whether exploitation is confirmed. 4. AUDIT: Review OAuth grant and access logs in connected platforms (Google Workspace, GitHub, Slack, etc.) for anomalous access in the past 30 days. 5. INVENTORY: Enumerate all Flowise instances in the environment including developer sandboxes and shadow IT deployments. 6. DETECT: Alert on unauthenticated GET requests to Flowise chatflow configuration endpoints in web/application logs.

What systems are affected by CVE-2026-41273?

This vulnerability affects the following AI/ML architecture patterns: agent frameworks, AI workflow automation, OAuth-integrated AI pipelines, enterprise AI orchestration, RAG pipelines with external data connectors.

What is the CVSS score for CVE-2026-41273?

CVE-2026-41273 has a CVSS v3.1 base score of 8.2 (HIGH). The EPSS exploitation probability is 0.31%.

What is the AI security impact?

Affected AI Architectures

agent frameworksAI workflow automationOAuth-integrated AI pipelinesenterprise AI orchestrationRAG pipelines with external data connectors

MITRE ATLAS Techniques

AML.T0049 Exploit Public-Facing Application

AML.T0083 Credentials from AI Agent Configuration

AML.T0084 Discover AI Agent Configuration

AML.T0091.000 Application Access Token

AML.T0106 Exploitation for Credential Access

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.1, A.7.4

NIST AI RMF: MANAGE 2.2

OWASP LLM Top 10: LLM02

What are the technical details?

Original Advisory

Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.0, Flowise contains an authentication bypass vulnerability that allows an unauthenticated attacker to obtain OAuth 2.0 access tokens associated with a public chatflow. By accessing a public chatflow configuration endpoint, an attacker can retrieve internal workflow data, including OAuth credential identifiers, which can then be used to refresh and obtain valid OAuth 2.0 access tokens without authentication. This vulnerability is fixed in 3.1.0.

Exploitation Scenario

An attacker scans internet-facing hosts for Flowise instances using Shodan or Censys queries targeting Flowise's distinctive UI fingerprint. They identify a target running a pre-3.1.0 version with public chatflows enabled. With a single unauthenticated HTTP GET to /api/v1/public-chatflows/{chatflow_id}, the attacker receives the full workflow configuration JSON, which includes OAuth credential identifiers for connected integrations — for example, a Google Drive connector used to feed documents into a RAG pipeline. The attacker passes these credential identifiers to the OAuth token refresh endpoint, obtaining a valid access token for the victim's Google Workspace account. From there, they exfiltrate documents from Google Drive, pivot to connected services using the same OAuth grant, and establish persistence by using the long-lived refresh token — all without ever authenticating to Flowise itself.

Weaknesses (CWE)

CWE-306 Missing Authentication for Critical Function Primary

CWE-306 — Missing Authentication for Critical Function: The product does not perform any authentication for functionality that requires a provable user identity or consumes a significant amount of resources.

[Architecture and Design] Divide the software into anonymous, normal, privileged, and administrative areas. Identify which of these areas require a proven user identity, and use a centralized authentication capability. Identify all potential communication channels, or other means of interaction with the software, to ensure that all channels are appropriately protected, including those channels that are assumed to be accessible only by authorized parties. Developers sometimes perform authentication at the primary channel, but open up a secondary channel that is assumed to be private. For example, a login mechanism may be listening on one network port, but after successful authentication, it may open up a second port where it waits for the connection, but avoids authentication because it assumes that only the authenticated party will connect to the port. In general, if the software or protocol allows a single session or user state to persist across multiple connections or channels, authentication and appropriate
[Architecture and Design] For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.

Source: MITRE CWE corpus.