CVE-2026-70477
CriticalCVSS 9.8Exploitation Probability (EPSS)
Elevated risk56th percentile - higher than 56% of all known CVEs
Summary
Flowise before version 3.1.3 contains a prompt injection vulnerability in the CSV Agent node, allowing a malicious Python script to bypass the blocklist validator and execute in an unsandboxed Pyodide environment. An attacker can leverage this to execute arbitrary code in the context of the service account.
Risk Assessment
The risk includes full compromise of the Flowise application and potentially the underlying server, leading to potential data confidentiality and integrity breaches.
Recommendation
It is recommended to immediately upgrade Flowise to version 3.1.3 or later, which includes a fix for this vulnerability.
Other vulnerabilities in Flowise
See all- CVE-2026-91930High
Flowise before 3.1.4 fails to scope enterprise organization and workspace membership APIs to the caller's tenant, allowing authenticated users to supply arbitrary organization IDs. An attacker can add themselves as organization owners, create workspaces, and gain administrative access to victim organizations.
- CVE-2026-91929High
Flowise versions before 3.1.4 contain cross-tenant authorization gaps in Enterprise endpoints that fail to verify resource ownership before operations. An attacker with Enterprise access can delete arbitrary workspaces, invite themselves into other organizations, modify cross-org roles, and abuse stored SSO secrets.
- CVE-2026-52098Critical
An issue in Flowise 3.1.2 allows a remote attacker to execute arbitrary code via the /api/v1/prediction/<flowId> endpoint.
- CVE-2026-73602Critical
Flowise before 3.1.3 contains a sandbox escape vulnerability in the vm2 JavaScript sandbox that allows authenticated users to execute arbitrary code by exploiting moment locale validation bypass.
- CVE-2026-73487Critical
Flowise before 3.1.3 contains a regex-based Python code validator bypass in CSV and Airtable Agent nodes that allows unauthenticated attackers to inject malicious code via prompt injection.
- CVE-2026-67622Critical
A vulnerability in Flowise (up to 3.1.4) in the OpenAI Assistants integration allows authenticated attackers to access credentials belonging to other workspaces by supplying an arbitrary credential UUID without workspace ownership verification.
- CVE-2026-70478Critical
Flowise prior to 3.1.3 contains a vulnerability in the POST /api/v1/oauth2-credential/refresh/:credentialId endpoint, which is in WHITELIST_URLS and requires no authentication. An attacker with a credential ID can refresh the OAuth token and gain access to the victim's connected service, as well as exhaust refresh-token quota.
- CVE-2026-70470Critical
Flowise prior to 3.1.3 contains a vulnerability in the Python code validator that can be bypassed with Unicode homoglyphs, allowing arbitrary Python execution inside Pyodide and full OS command execution on the Flowise host.
- CVE-2026-69264Critical
Flowise prior to 3.1.3 contains a vulnerability in CSVAgent that allows an attacker to inject code into a Python template executed by Pyodide, enabling arbitrary OS command execution on the Flowise host. An unauthenticated request to the API can trigger remote code execution.
- CVE-2026-69263Critical
Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.3, the mitigation for CVE-2025-8943 blocked -y and --yes flags on npx, but packages/components/nodes/tools/MCP/core.ts denied only PATH, LD_LIBRARY_PATH, DYLD_LIBRARY_PATH, and NODE_OPTIONS by exact environment-variable name. Because npm reads configuration from npm_config_* variables, setting npm_config_yes=true reproduced --yes behavior without using a blocked flag, causing npx to auto-install and execute the named package when a Custom MCP server launched. This issue is fixed in version 3.1.3.
Original NVD description (English source)
Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.3, a prompt injection sent to a chatflow using a CSV Agent node can cause the LLM to respond with a malicious Python script that bypasses the blocklist validator and executes in an unsandboxed Pyodide environment. The specific flaw exists within the run method of the CSV_Agents class, where untrusted data is used to construct an LLM prompt and the resulting pythonCode is validated by validatePythonCodeForDataFrame before execution. An attacker can leverage this to execute arbitrary code in the context of the service account. This issue is fixed in 3.1.3.