CVE-2026-130796: Flowise: public chatflow leaks MCP bearer token

HIGH
Published November 17, 2026
CISO Take

Flowise 3.1.3 through 3.1.4 exposes MCP server bearer tokens to anyone, with no login, through the public-chatflows endpoint. An attacker can reuse the leaked token on the MCP route to run chatflows directly, which bypasses API key protection and burns the owner's LLM quota and any data or tools the flow can reach. No EPSS score, public exploit, Nuclei template or KEV listing exists yet, but the attack is a simple two-request sequence over the network and the reference repositories describe the flaw in detail, so weaponization is likely to be quick. Upgrade beyond 3.1.4 where a fix is available (the vendor sunset notice should be reviewed for support status), rotate all MCP tokens on exposed chatflows, and unpublish public chatflows that don't need to be public.

Sources:NVDATLASvulncheck.comgithub.com/FlowiseAI/Flowise

What is the risk?

High (CVSS 8.2). Exploitation needs no authentication, no user interaction and low complexity over the network. Confidentiality impact is high because bearer tokens are disclosed; integrity impact is low because the token lets an attacker invoke chatflows and trigger their side effects. No active exploitation, public exploit or scanner template is known today, and the EPSS score is unavailable. Exposure is highest for internet-facing Flowise instances with public chatflows that have MCP server configuration enabled. Flowise is a low-code agent builder, so a chatflow may carry connected credentials, tools and knowledge bases, which raises the potential impact beyond quota theft.

How does the attack unfold?

Discovery
Attacker finds an internet-facing Flowise instance and a public chatflow ID.
AML.T0006
Token Disclosure
Unauthenticated GET to /api/v1/public-chatflows/:id returns the mcpServerConfig bearer token.
AML.T0049
Authentication Bypass
Attacker presents the leaked token on POST /api/v1/mcp/:chatflowId, skipping API key protection.
AML.T0091.000
Impact
Attacker repeatedly invokes the chatflow, consuming model quota and potentially abusing connected tools.
AML.T0034

What systems are affected?

PackageEcosystemVulnerable RangePatched
Flowisenpm—No patch

Do you use Flowise? You're affected.

How severe is it?

CVSS 3.1
8.2 / 10
EPSS
N/A
Exploitation Status
No known exploitation
Sophistication
Trivial

What is the attack surface?

AVACPRUISCIA
AVNetwork
ACLow
PRNone
UINone
SUnchanged
CHigh
ILow
ANone

What should I do?

1 step
  1. 1) Upgrade Flowise to a version later than 3.1.4 that fixes the token exposure, and review the vendor sunset notice (flowiseai.com/sunset) for support and migration implications. 2) Rotate all MCP server bearer tokens on every chatflow that was public while running 3.1.3-3.1.4, and treat existing tokens as compromised. 3) Until patched, set affected chatflows to non-public or disable MCP server configuration, and restrict access to /api/v1/public-chatflows and /api/v1/mcp at the reverse proxy or WAF. 4) Detect abuse by reviewing access logs for unauthenticated GET requests to /api/v1/public-chatflows/:id followed by POST requests to /api/v1/mcp/:chatflowId from the same source, and for spikes in token usage or model spend. 5) Set provider-side spend limits and alerts on the LLM keys used by Flowise.

How is it classified?

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act
Art. 15-Accuracy, robustness and cybersecurity
ISO 42001
A.6.2.6-AI system operation and monitoring
NIST AI RMF
MANAGE 2.3-Respond to previously unknown risks
OWASP LLM Top 10
LLM06-Excessive AgencyLLM10-Unbounded Consumption

Frequently Asked Questions

What is CVE-2026-130796?

Flowise 3.1.3 through 3.1.4 exposes MCP server bearer tokens to anyone, with no login, through the public-chatflows endpoint. An attacker can reuse the leaked token on the MCP route to run chatflows directly, which bypasses API key protection and burns the owner's LLM quota and any data or tools the flow can reach. No EPSS score, public exploit, Nuclei template or KEV listing exists yet, but the attack is a simple two-request sequence over the network and the reference repositories describe the flaw in detail, so weaponization is likely to be quick. Upgrade beyond 3.1.4 where a fix is available (the vendor sunset notice should be reviewed for support status), rotate all MCP tokens on exposed chatflows, and unpublish public chatflows that don't need to be public.

Is CVE-2026-130796 actively exploited?

No confirmed active exploitation of CVE-2026-130796 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-130796?

1) Upgrade Flowise to a version later than 3.1.4 that fixes the token exposure, and review the vendor sunset notice (flowiseai.com/sunset) for support and migration implications. 2) Rotate all MCP server bearer tokens on every chatflow that was public while running 3.1.3-3.1.4, and treat existing tokens as compromised. 3) Until patched, set affected chatflows to non-public or disable MCP server configuration, and restrict access to /api/v1/public-chatflows and /api/v1/mcp at the reverse proxy or WAF. 4) Detect abuse by reviewing access logs for unauthenticated GET requests to /api/v1/public-chatflows/:id followed by POST requests to /api/v1/mcp/:chatflowId from the same source, and for spikes in token usage or model spend. 5) Set provider-side spend limits and alerts on the LLM keys used by Flowise.

What systems are affected by CVE-2026-130796?

This vulnerability affects the following AI/ML architecture patterns: Agent frameworks, Low-code LLM workflow builders, MCP server deployments, Public chatbot deployments.

What is the CVSS score for CVE-2026-130796?

CVE-2026-130796 has a CVSS v3.1 base score of 8.2 (HIGH).

What is the AI security impact?

Affected AI Architectures

Agent frameworksLow-code LLM workflow buildersMCP server deploymentsPublic chatbot deployments

MITRE ATLAS Techniques

AML.T0034Cost Harvesting
AML.T0049Exploit Public-Facing Application
AML.T0053AI Agent Tool Invocation
AML.T0091.000Application Access Token
AML.T0096AI Service API

Compliance Controls Affected

EU AI Act: Art. 15
ISO 42001: A.6.2.6
NIST AI RMF: MANAGE 2.3
OWASP LLM Top 10: LLM06, LLM10

What are the technical details?

Original Advisory

Flowise 3.1.3 through 3.1.4 contains an information disclosure vulnerability that allows unauthenticated attackers to read MCP server bearer tokens via the GET /api/v1/public-chatflows/:id handler. Attackers can use the leaked mcpServerConfig token on the POST /api/v1/mcp/:chatflowId route to invoke chatflows, bypassing API key protection and consuming the owner's model quota.

Exploitation Scenario

An attacker scans the internet for Flowise instances and finds a public chatflow ID, which is often visible in embedded chat widgets or shared links. They send an unauthenticated GET to /api/v1/public-chatflows/:id and read the mcpServerConfig bearer token from the JSON response. They then POST to /api/v1/mcp/:chatflowId with that token to invoke the chatflow repeatedly, which skips API key checks. The attacker can resell the free LLM access, drive up the owner's bill, or feed crafted inputs through the flow's connected tools and data sources to extract information or trigger actions.

Weaknesses (CWE)

CWE-212 — Improper Removal of Sensitive Information Before Storage or Transfer: The product stores, transfers, or shares a resource that contains sensitive information, but it does not properly remove that information before the product makes the resource available to unauthorized actors.

  • [Requirements] Clearly specify which information should be regarded as private or sensitive, and require that the product offers functionality that allows the user to cleanse the sensitive information from the resource before it is published or exported to other parties.
  • [Architecture and Design] Compartmentalize the system to have "safe" areas where trust boundaries can be unambiguously drawn. Do not allow sensitive data to go outside of the trust boundary and always be careful when interfacing with a compartment outside of the safe area. Ensure that appropriate compartmentalization is built into the system design, and the compartmentalization allows for and reinforces privilege separation functionality. Architects and designers should rely on the principle of least privilege to decide the appropriate time to use privileges and the time to drop privileges.

Source: MITRE CWE corpus.

CVSS Vector

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N

Timeline

Published
November 17, 2026
Last Modified
November 17, 2026
First Seen
November 17, 2026

Related Vulnerabilities