CVE-2025-68478: CWE-73: External Control of File Name or Path in langflow-ai langflow
Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.7.0, if an arbitrary path is specified in the request body's `fs_path`, the server serializes the Flow object into JSON and creates/overwrites a file at that path. There is no path restriction, normalization, or allowed directory enforcement, so absolute paths (e.g., /etc/poc.txt) are interpreted as is. Version 1.7.0 fixes the issue.
AI Analysis
Technical Summary
CVE-2025-68478 is a vulnerability classified under CWE-73 (External Control of File Name or Path) affecting langflow, an AI workflow and agent building tool. In versions prior to 1.7.0, the application accepts a parameter named fs_path in the request body, which specifies the file system path where the serialized Flow object JSON is saved. Due to lack of path normalization, restriction, or enforcement of allowed directories, an attacker can supply arbitrary absolute paths, causing the server to create or overwrite files anywhere on the filesystem accessible by the application process. This can lead to unauthorized modification of critical files, potentially altering application behavior or system configurations, thus compromising integrity. The vulnerability requires network access and low privileges (PR:L), but no user interaction is needed. The CVSS v3.1 score is 7.1 (high), reflecting the ease of exploitation and the significant impact on integrity with some availability impact. No known exploits are reported in the wild yet. The issue was addressed in langflow version 1.7.0 by implementing proper path validation and restrictions to prevent arbitrary file writes outside intended directories.
Potential Impact
For European organizations, this vulnerability poses a significant risk to the integrity of systems running vulnerable langflow versions. Attackers could overwrite configuration files, inject malicious code, or disrupt AI workflow operations by tampering with serialized flow files. This could lead to operational disruptions, loss of trust in AI-driven processes, or further compromise if attackers leverage file overwrites to escalate privileges or pivot within networks. Although confidentiality is not directly impacted, the integrity and availability of critical AI workflows and related services could be degraded. Organizations in sectors relying heavily on AI automation, such as finance, manufacturing, and research institutions, may face operational and reputational damage. The vulnerability's network accessibility and lack of user interaction requirement increase the risk of remote exploitation, making timely patching essential.
Mitigation Recommendations
European organizations should immediately upgrade langflow installations to version 1.7.0 or later, where the vulnerability is fixed. If immediate upgrade is not feasible, implement strict network segmentation and firewall rules to restrict access to langflow services only to trusted internal users and systems. Monitor logs for unusual file write operations or unexpected paths in requests to detect exploitation attempts. Employ application-layer controls or web application firewalls (WAFs) to validate and sanitize input parameters, specifically fs_path, to prevent arbitrary path injection. Conduct regular audits of file system integrity and implement file system permissions to limit the application’s write access to only necessary directories. Additionally, educate developers and administrators about secure coding practices related to file path handling to prevent similar issues in custom workflows.
Affected Countries
Germany, France, United Kingdom, Netherlands, Sweden, Finland
CVE-2025-68478: CWE-73: External Control of File Name or Path in langflow-ai langflow
Description
Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.7.0, if an arbitrary path is specified in the request body's `fs_path`, the server serializes the Flow object into JSON and creates/overwrites a file at that path. There is no path restriction, normalization, or allowed directory enforcement, so absolute paths (e.g., /etc/poc.txt) are interpreted as is. Version 1.7.0 fixes the issue.
AI-Powered Analysis
Technical Analysis
CVE-2025-68478 is a vulnerability classified under CWE-73 (External Control of File Name or Path) affecting langflow, an AI workflow and agent building tool. In versions prior to 1.7.0, the application accepts a parameter named fs_path in the request body, which specifies the file system path where the serialized Flow object JSON is saved. Due to lack of path normalization, restriction, or enforcement of allowed directories, an attacker can supply arbitrary absolute paths, causing the server to create or overwrite files anywhere on the filesystem accessible by the application process. This can lead to unauthorized modification of critical files, potentially altering application behavior or system configurations, thus compromising integrity. The vulnerability requires network access and low privileges (PR:L), but no user interaction is needed. The CVSS v3.1 score is 7.1 (high), reflecting the ease of exploitation and the significant impact on integrity with some availability impact. No known exploits are reported in the wild yet. The issue was addressed in langflow version 1.7.0 by implementing proper path validation and restrictions to prevent arbitrary file writes outside intended directories.
Potential Impact
For European organizations, this vulnerability poses a significant risk to the integrity of systems running vulnerable langflow versions. Attackers could overwrite configuration files, inject malicious code, or disrupt AI workflow operations by tampering with serialized flow files. This could lead to operational disruptions, loss of trust in AI-driven processes, or further compromise if attackers leverage file overwrites to escalate privileges or pivot within networks. Although confidentiality is not directly impacted, the integrity and availability of critical AI workflows and related services could be degraded. Organizations in sectors relying heavily on AI automation, such as finance, manufacturing, and research institutions, may face operational and reputational damage. The vulnerability's network accessibility and lack of user interaction requirement increase the risk of remote exploitation, making timely patching essential.
Mitigation Recommendations
European organizations should immediately upgrade langflow installations to version 1.7.0 or later, where the vulnerability is fixed. If immediate upgrade is not feasible, implement strict network segmentation and firewall rules to restrict access to langflow services only to trusted internal users and systems. Monitor logs for unusual file write operations or unexpected paths in requests to detect exploitation attempts. Employ application-layer controls or web application firewalls (WAFs) to validate and sanitize input parameters, specifically fs_path, to prevent arbitrary path injection. Conduct regular audits of file system integrity and implement file system permissions to limit the application’s write access to only necessary directories. Additionally, educate developers and administrators about secure coding practices related to file path handling to prevent similar issues in custom workflows.
Affected Countries
For access to advanced analysis and higher rate limits, contact root@offseq.com
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- GitHub_M
- Date Reserved
- 2025-12-18T18:29:07.309Z
- Cvss Version
- 3.1
- State
- PUBLISHED
Threat ID: 69458ae3f063e4fadf03d17a
Added to database: 12/19/2025, 5:26:59 PM
Last enriched: 12/19/2025, 5:41:55 PM
Last updated: 12/19/2025, 8:01:16 PM
Views: 74
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Related Threats
CVE-2025-12874: CWE-444 Inconsistent Interpretation of HTTP Requests ('HTTP Request/Response Smuggling') in Quest Coexistence Manager for Notes
MediumCVE-2025-14967: SQL Injection in itsourcecode Student Management System
MediumCVE-2025-14966: SQL Injection in FastAdmin
MediumCVE-2025-14964: Stack-based Buffer Overflow in TOTOLINK T10
CriticalCVE-2025-14965: Path Traversal in 1541492390c yougou-mall
MediumActions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
External Links
Need enhanced features?
Contact root@offseq.com for Pro access with improved analysis and higher rate limits.