AI Threat AlertCVE-2026-97677: Langflow: path traversal enables file write and read
HIGHLangflow OSS (IBM-distributed, 1.0.0 through 1.12.2) has a path traversal flaw (CWE-22) that lets any authenticated flow author write attacker-controlled content to any directory the service account can write to, and read arbitrary files via a crafted on-disk index. This bypasses the platform's local file access isolation, so configuration files, secrets, API keys for LLM providers and database files in the application data directory are exposed. Exploitation needs only low privileges and no user interaction (CVSS 8.1), but there is no CISA KEV listing, no public exploit, no Nuclei template and no EPSS score yet, so there is no evidence of active exploitation at publication. Langflow has a history of high-impact bugs and Metasploit coverage for related flaws, so expect weaponization to be quick. Follow the IBM advisory (ibm.com/support/pages/node/7290694) and upgrade past 1.12.2 to the fixed release. Until then, restrict flow-author accounts, run the service as a low-privilege user with a minimal writable filesystem, and rotate any secrets stored under the data directory if untrusted users had access.
What is the risk?
High. The network attack vector, low complexity and low privilege requirement make exploitation straightforward for anyone with a flow-author account, including a compromised or malicious insider. Confidentiality and integrity impact are both high (C:H/I:H), while availability is unaffected. Exposure depends on who can author flows; multi-tenant or internet-facing Langflow instances with open signup carry the greatest risk. No active exploitation, public exploit code or EPSS data is available, so urgency is high but not emergency-level. The arbitrary file write primitive could be chained to code execution (for example by overwriting scripts, configs or startup files), which would raise the effective impact beyond the scored rating.
How does the attack unfold?
What systems are affected?
| Package | Ecosystem | Vulnerable Range | Patched |
|---|---|---|---|
| Langflow | pip | — | No patch |
Do you use Langflow? You're affected.
How severe is it?
What is the attack surface?
What should I do?
1 step1) Upgrade IBM Langflow OSS beyond 1.12.2 to the fixed version listed in IBM advisory 7290694. 2) Until patched, limit flow-author privileges to trusted users, disable open self-registration and put the instance behind SSO/VPN. 3) Run Langflow in a container with a read-only root filesystem, a dedicated low-privilege user and only the necessary writable volumes. 4) Move secrets out of the data directory into a secrets manager and rotate LLM API keys, database credentials and the Langflow secret key if exposure is suspected. 5) Detect: monitor file writes outside expected paths by the Langflow process, flow components or API calls containing '../' or absolute paths, and unexpected reads of config or database files. 6) Review audit logs for unusual flow creation or modification by low-privileged accounts.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-97677?
Langflow OSS (IBM-distributed, 1.0.0 through 1.12.2) has a path traversal flaw (CWE-22) that lets any authenticated flow author write attacker-controlled content to any directory the service account can write to, and read arbitrary files via a crafted on-disk index. This bypasses the platform's local file access isolation, so configuration files, secrets, API keys for LLM providers and database files in the application data directory are exposed. Exploitation needs only low privileges and no user interaction (CVSS 8.1), but there is no CISA KEV listing, no public exploit, no Nuclei template and no EPSS score yet, so there is no evidence of active exploitation at publication. Langflow has a history of high-impact bugs and Metasploit coverage for related flaws, so expect weaponization to be quick. Follow the IBM advisory (ibm.com/support/pages/node/7290694) and upgrade past 1.12.2 to the fixed release. Until then, restrict flow-author accounts, run the service as a low-privilege user with a minimal writable filesystem, and rotate any secrets stored under the data directory if untrusted users had access.
Is CVE-2026-97677 actively exploited?
No confirmed active exploitation of CVE-2026-97677 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-97677?
1) Upgrade IBM Langflow OSS beyond 1.12.2 to the fixed version listed in IBM advisory 7290694. 2) Until patched, limit flow-author privileges to trusted users, disable open self-registration and put the instance behind SSO/VPN. 3) Run Langflow in a container with a read-only root filesystem, a dedicated low-privilege user and only the necessary writable volumes. 4) Move secrets out of the data directory into a secrets manager and rotate LLM API keys, database credentials and the Langflow secret key if exposure is suspected. 5) Detect: monitor file writes outside expected paths by the Langflow process, flow components or API calls containing '../' or absolute paths, and unexpected reads of config or database files. 6) Review audit logs for unusual flow creation or modification by low-privileged accounts.
What systems are affected by CVE-2026-97677?
This vulnerability affects the following AI/ML architecture patterns: Agent frameworks, LLM workflow orchestration, RAG pipelines, Multi-tenant AI development platforms.
What is the CVSS score for CVE-2026-97677?
CVE-2026-97677 has a CVSS v3.1 base score of 8.1 (HIGH). The EPSS exploitation probability is 0.39%.
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0037Data from Local SystemAML.T0049Exploit Public-Facing ApplicationAML.T0055Unsecured CredentialsAML.T0083Credentials from AI Agent ConfigurationCompliance Controls Affected
What are the technical details?
Original Advisory
IBM Langflow OSS 1.0.0 through 1.12.2 An authenticated flow author could write attacker-controlled file content into any directory writable by the service account, bypassing the platform's local file access isolation safeguard. Via a crafted on-disk index, the attacker could also read the contents of arbitrary files accessible to the service process — including configuration files, secrets, or database files stored under the application's data directory.
Exploitation Scenario
An attacker obtains or already holds a flow-author account on a shared Langflow server. They build a flow using a file-handling component and supply a crafted path with traversal sequences, writing a malicious file into a writable directory such as a startup script location or a web-served path. They then craft an on-disk index that points to sensitive files and trigger a read, retrieving the .env file, the SQLite database and stored LLM provider API keys. With these credentials they pivot to the organization's cloud AI accounts and vector databases, exfiltrating proprietary data and running up usage costs.
Weaknesses (CWE)
CWE-22Improper Limitation of a Pathname to a Restricted DirectoryPrimaryCWE-22Improper Limitation of a Pathname to a Restricted DirectoryCWE-22 — Improper Limitation of a Pathname to a Restricted Directory: 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.
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:NReferences
Timeline
Related Vulnerabilities
CVE-2026-1056110.0Langflow: auth bypass + unauthenticated RCE (CVSS 10)
Same package: langflowCVE-2026-1013410.0Langflow: unauthenticated RCE via tool_code injection
Same package: langflowCVE-2026-333099.9langflow: Path Traversal enables file access
Same package: langflowCVE-2026-552559.9Langflow: IDOR allows cross-user flow execution
Same package: langflowCVE-2026-78739.9Langflow: authenticated RCE enables credential theft
Same package: langflow