CVE-2026-9196
HighCVSS 8.1Exploitation Probability (EPSS)
Low risk18th percentile - higher than 18% of all known CVEs
Summary
IBM Langflow OSS versions 1.0.0 through 1.10.3 allow an authenticated attacker to execute unintended code during Agentic Assistant validation due to improper handling of LLM-generated components. The application executes model-generated Python code in the backend before user approval, which may lead to side effects such as network access, file system interaction, or data exfiltration.
Risk Assessment
The risk includes the possibility of unauthorized actions on the system, including data leakage or system integrity compromise, with the privileges of the Langflow backend process.
Recommendation
Upgrade to a version later than 1.10.3 that contains a fix, and implement approval mechanisms before executing LLM-generated code.
Other vulnerabilities in IBM Langflow
See all- CVE-2026-8859Critical
IBM Langflow OSS 1.0.0 through 1.10.0 contains a path traversal vulnerability in the APIRequest component, where filenames from Content-Disposition headers are not sanitized. An attacker can write arbitrary files to locations accessible by the Langflow process.
- CVE-2026-9202Critical
IBM Langflow OSS versions 1.0.0 through 1.10.0 allow unauthenticated attackers to create unlimited user accounts. When NEW_USER_IS_ACTIVE=true is set, new accounts are immediately active and can access RCE endpoints.
- CVE-2026-9205High
IBM Langflow OSS contains a weakness in the ensure_fernet_key() function related to weak cryptographic key derivation.
- CVE-2026-9201High
IBM Langflow OSS versions 1.0.0 through 1.10.3 allow an authenticated attacker to execute arbitrary code due to a cryptographic weakness in the custom component validation mechanism. Validation uses a truncated SHA-256 hash, enabling collisions and bypassing validation, leading to execution of unauthorized Python code.
- CVE-2026-10128Medium
IBM Langflow OSS versions 1.0.0 through 1.10.3 allow authenticated users to exploit a built-in component to read arbitrary server environment variables. This exposes sensitive secrets despite security controls intended to disable custom components.
- CVE-2026-7646Medium
IBM Langflow OSS 1.0.0 through 1.10.3 allows users to read arbitrary files from the server filesystem, including other users' uploaded documents, the JWT signing secret, the SQLite database, and process environment variables, by sending a crafted MCP `resources/read` request with a URL-encoded path traversal sequence in the filename.
- CVE-2026-17630High
A vulnerability in IBM Langflow OSS versions 1.0.0 through 1.10.3 allows a remote attacker to execute arbitrary code due to improper validation of configuration parameters.
- CVE-2026-17626High
IBM Langflow OSS versions 1.0.0 through 1.10.3 allow an authenticated attacker to read, modify, or expose sensitive host files via Docker-based MCP servers due to incomplete filtering of dangerous Docker volume-mount and device-mapping arguments.
- CVE-2026-17623High
IBM Langflow OSS versions 1.0.0 through 1.10.3 allow a remote authenticated attacker to execute arbitrary commands due to improper validation of the command field in MCP server configurations.
- CVE-2026-10700Medium
IBM Langflow OSS has multiple broken access control vulnerabilities in its file handling API. Unauthenticated attackers can retrieve image files, and authenticated users can access files belonging to other users.
Original NVD description (English source)
IBM Langflow OSS 1.0.0 through 1.10.3 could allow an authenticated attacker to execute unintended code during Agentic Assistant validation due to improper handling of LLM‑generated components. The application executes model‑generated Python code in the backend during validation prior to user approval, which may allow an attacker to trigger side effects such as outbound network access, file system interaction, or data exfiltration with the privileges of the Langflow backend process.