CVE-2026-56273: Flowise path traversal

CISO Take

Flowise, an open-source low-code platform for building AI agents and LLM workflows, fails to sanitize the basePath parameter accepted by its Faiss and SimpleStore vector store connectors, letting any authenticated API-token holder redirect vector store writes to arbitrary filesystem paths — a classic CWE-22 path traversal rated CVSS 6.5 (AV:N/AC:L/PR:L/UI:N, integrity-high impact, no confidentiality or availability loss). There's no EPSS score, no CISA KEV listing, and no public exploit or Nuclei template yet, so this isn't an active-exploitation emergency, but the network-reachable, low-complexity, no-user-interaction profile means a malicious or compromised low-privilege API user could weaponize it quickly, and arbitrary file writes on a server hosting an AI agent orchestrator are a well-worn path to remote code execution. Organizations self-hosting Flowise for agent workflows or RAG pipelines should treat this as a priority patch, since Flowise instances are frequently deployed with broad filesystem and credential access for their agent tools. Upgrade to Flowise 3.1.0 or later immediately; until patched, restrict API token issuance, audit basePath values in vector store configs for traversal sequences, and monitor for file writes outside expected data directories on hosts running Flowise.

Sources: NVD GitHub Advisory vulncheck.com ATLAS

What is the risk?

Medium overall risk: exploitability is high on paper (network vector, low attack complexity, no user interaction) but gated by a real precondition — the attacker needs a valid, if low-privileged, API token, which narrows the population of attackers to insiders, compromised integrations, or holders of leaked/shared tokens. Impact is integrity-only per the CVSS vector (no direct confidentiality or availability loss), but arbitrary filesystem write on a server that also runs an AI agent orchestration engine is a strong stepping stone to full code execution — the practical severity is higher than the raw score suggests. No EPSS data, no KEV listing, and no known public exploit or scanner template mean there is currently no evidence of in-the-wild exploitation, but the vulnerability class (unsanitized path parameter) is trivial to exploit once an attacker has the required token.

How does the attack unfold?

Authenticated access

Attacker obtains or already holds a valid, low-privilege Flowise API token.

AML.T0012

Path traversal exploitation

Attacker configures a Faiss or SimpleStore vector store with a basePath containing directory traversal sequences.

AML.T0049

Arbitrary file write

Flowise writes vector store data to the attacker-chosen filesystem path instead of the intended data directory.

Escalation to code execution

If the written file lands in an executed or trusted location (web root, cron job, SSH keys), the attacker achieves remote code execution or persistent access on the host.

AML.T0072

Authenticated access

Attacker obtains or already holds a valid, low-privilege Flowise API token.

AML.T0012

Path traversal exploitation

Attacker configures a Faiss or SimpleStore vector store with a basePath containing directory traversal sequences.

AML.T0049

Arbitrary file write

Flowise writes vector store data to the attacker-chosen filesystem path instead of the intended data directory.

Escalation to code execution

If the written file lands in an executed or trusted location (web root, cron job, SSH keys), the attacker achieves remote code execution or persistent access on the host.

AML.T0072

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

6.5 / 10

EPSS

0.3%

chance of exploitation in 30 days

Higher than 26% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Available

Exploitation: MEDIUM

Sophistication

Moderate

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 Low

UI None

S Unchanged

C None

I High

A None

What should I do?

1 step

Upgrade Flowise to 3.1.0 or later immediately, which sanitizes/validates the basePath parameter. Until patched: restrict and rotate API tokens, limiting who can configure Faiss/SimpleStore vector stores; run Flowise under a least-privilege OS user with a chrooted or containerized filesystem so path traversal cannot reach sensitive host paths; add input validation/WAF rules that reject basePath values containing ../, absolute paths, or symlink indicators; and monitor filesystem write activity from the Flowise process for writes outside its configured data directory. Review existing vector store configs for suspicious basePath values as a compromise-assessment step.

What does CISA's SSVC say?

Decision Track*

Exploitation poc

Automatable No

Technical Impact partial

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

How is it classified?

Code Execution Framework Agent RAG AML.T0012 - Valid Accounts AML.T0049 - Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Article 15 - Accuracy, robustness and cybersecurity

ISO 42001

A.6.2.3 - System security for AI systems

NIST AI RMF

MEASURE 2.7 - AI system security and resilience is evaluated and documented

OWASP LLM Top 10

LLM07 - Insecure Plugin Design

Frequently Asked Questions

What is CVE-2026-56273?

Flowise, an open-source low-code platform for building AI agents and LLM workflows, fails to sanitize the basePath parameter accepted by its Faiss and SimpleStore vector store connectors, letting any authenticated API-token holder redirect vector store writes to arbitrary filesystem paths — a classic CWE-22 path traversal rated CVSS 6.5 (AV:N/AC:L/PR:L/UI:N, integrity-high impact, no confidentiality or availability loss). There's no EPSS score, no CISA KEV listing, and no public exploit or Nuclei template yet, so this isn't an active-exploitation emergency, but the network-reachable, low-complexity, no-user-interaction profile means a malicious or compromised low-privilege API user could weaponize it quickly, and arbitrary file writes on a server hosting an AI agent orchestrator are a well-worn path to remote code execution. Organizations self-hosting Flowise for agent workflows or RAG pipelines should treat this as a priority patch, since Flowise instances are frequently deployed with broad filesystem and credential access for their agent tools. Upgrade to Flowise 3.1.0 or later immediately; until patched, restrict API token issuance, audit basePath values in vector store configs for traversal sequences, and monitor for file writes outside expected data directories on hosts running Flowise.

Is CVE-2026-56273 actively exploited?

Proof-of-concept exploit code is publicly available for CVE-2026-56273, increasing the risk of exploitation.

How to fix CVE-2026-56273?

Upgrade Flowise to 3.1.0 or later immediately, which sanitizes/validates the basePath parameter. Until patched: restrict and rotate API tokens, limiting who can configure Faiss/SimpleStore vector stores; run Flowise under a least-privilege OS user with a chrooted or containerized filesystem so path traversal cannot reach sensitive host paths; add input validation/WAF rules that reject basePath values containing `../`, absolute paths, or symlink indicators; and monitor filesystem write activity from the Flowise process for writes outside its configured data directory. Review existing vector store configs for suspicious basePath values as a compromise-assessment step.

What systems are affected by CVE-2026-56273?

This vulnerability affects the following AI/ML architecture patterns: RAG pipelines, agent frameworks, vector databases, self-hosted AI orchestration platforms.

What is the CVSS score for CVE-2026-56273?

CVE-2026-56273 has a CVSS v3.1 base score of 6.5 (MEDIUM). The EPSS exploitation probability is 0.33%.

What is the AI security impact?

Affected AI Architectures

RAG pipelinesagent frameworksvector databasesself-hosted AI orchestration platforms

MITRE ATLAS Techniques

AML.T0012 Valid Accounts

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Article 15

ISO 42001: A.6.2.3

NIST AI RMF: MEASURE 2.7

OWASP LLM Top 10: LLM07

What are the technical details?

Original Advisory

Flowise before 3.1.0 contains a path traversal vulnerability in Faiss and SimpleStore vector store implementations that accept unsanitized basePath parameters from authenticated users. Attackers with valid API tokens can write vector store data to arbitrary filesystem locations, potentially enabling code execution or data exfiltration.

Exploitation Scenario

An attacker who holds (or steals/leaks) a low-privilege Flowise API token calls the API to create or update a Faiss or SimpleStore vector store, setting basePath to a traversal string such as `../../../../var/www/html/shell.php` or a path targeting a cron directory or SSH authorized_keys file. When Flowise persists the vector store data, it writes attacker-controlled content to that arbitrary location instead of the intended data directory. If the written file lands somewhere the host executes or trusts (a web root, a cron job, an init script), the attacker escalates from a data-write primitive to full remote code execution on the Flowise host, potentially pivoting into any connected RAG data sources, credentials, or downstream agent tools.

Weaknesses (CWE)

CWE-22 Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal') Primary CWE-22 Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal')

CWE-22 — Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'): The product uses external input to construct a pathname that is intended to identify a file or directory that is located underneath a restricted parent directory, but the product does not properly neutralize special elements within the pathname that can cause the pathname to resolve to a location that is outside of the restricted directory.

[Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
[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.