Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessReach exposed Langflow HTTP API
-
DeliverySend crafted create-knowledge-base POST
-
ExploitInject absolute path in request
-
ExecutionWrite file outside workspace boundary
-
ImpactOverwrite config or plant payload for code execution
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V5 - File Handling Path canonicalization, traversal prevention
MITRE ATT&CK
2 techniques across 2 tactics: Discovery, Collection The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Requires the attacker to reach the Langflow backend's knowledge-base creation route (HTTP POST to /api/v1/knowledge_bases, handler at knowledge_bases.py:51) and to supply an attacker-controlled absolute path/filename, which the handler writes to host storage without boundary enforcement. No special non-default feature toggle is named in the description. Limiting factors: the Langflow API must be network-reachable, and although CVSS rates this PR:N, Langflow installations that enforce API login would require valid authentication, so the practical prerequisite in many environments is an authenticated session plus an exposed backend; the overall impact (overwrite of files at a chosen absolute path) is bounded by the OS privileges of the Langflow process. |
| Impact | Signals are moderate and largely consistent. The NVD CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:N, score 7.5) describes a network-reachable, low-complexity flaw needing no privileges or user interaction, with high integrity impact and no confidentiality/availability impact - consistent with arbitrary file write via absolute-path traversal. EPSS is low (0.23%, 13th percentile) and the CVE is NOT in CISA KEV, so there is no evidence of active or widespread exploitation. A referenced public gist (Ro1ME) appears to be a proof-of-concept/write-up, so publicly available exploit code likely exists, but POC availability is a distinct signal from active exploitation and does not elevate this to in-the-wild status. CONFLICT TO VERIFY: CVSS asserts PR:N (unauthenticated), but Langflow normally ships with authentication on its API, so real-world exploitation may require authenticated access (PR:L) depending on deployment - treat the unauthenticated rating as the description-provided value, not an independently confirmed fact. No SSVC decision data was provided. Net: this is a real but second-tier priority - genuine arbitrary-file-write primitive that can escalate to code execution via overwritten config/code, tempered by low EPSS, no KEV, and likely authentication/exposure prerequisites in many deployments. |
| Exploit ScenarioDerived | An attacker with network access to an exposed Langflow backend sends an HTTP POST to the knowledge-base creation endpoint supplying an absolute path (e.g. a system config or startup script location) as the target. Because the handler forwards the attacker-controlled path into file creation without boundary enforcement, a file is written or overwritten outside the intended workspace, which can corrupt configuration or plant executable content leading to code execution. A public gist proof-of-concept is referenced, lowering the barrier to weaponization. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 06, 2026
Absolute path traversal in langflow-ai Langflow through v1.8.4 lets a remote attacker write or overwrite files anywhere the Langflow process can reach by supplying an attacker-controlled absolute path or filename to the knowledge-base creation endpoint (HTTP POST /api/v1/knowledge_bases, handler at src/backend/base/langflow/api/v1/knowledge_bases.py:51), which forwards that input into host file creation without enforcing a workspace or storage boundary. The CVSS 3.1 vector (7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:N) rates it unauthenticated with high integrity impact, although the independent assessment notes that installations enforcing API login make a valid session the practical prerequisite and that the blast radius is capped by the OS privileges of the Langflow service account. Exploitation pressure is currently low per EPSS (0.23%, 13th percentile), and no public exploit has been identified at time of analysis.
RemediationAI
No vendor-released patch was identified at time of analysis, so the first action is to confirm a fixed Langflow release from the upstream repository or vendor advisory and upgrade from v1.8.4 (or any earlier affected build) as soon as one is published, cross-checking the NVD entry at https://nvd.nist.gov/vuln/detail/CVE-2026-51888 and the disclosure gist at https://gist.github.com/Ro1ME/c85f513d1fd702c686cffe061cbf8ef5 since no exact fixed version can be cited from the available data. Until then, apply compensating controls that directly counter arbitrary out-of-workspace writes: run Langflow as a dedicated unprivileged account inside a container with a read-only root filesystem and narrowly scoped writable volumes, which causes writes outside the workspace to fail at the OS level (trade-off: legitimate features that persist data outside those mounts will break until mounts are tuned); restrict network reachability of the backend so the knowledge-base route is only exposed to trusted clients behind an authenticating reverse proxy, which is important because the CVSS vector is PR:N even though deployments with enforced API login effectively require a session (trade-off: unauthenticated API clients and integrations stop working); disable or block the knowledge-base creation endpoint entirely if the feature is unused (trade-off: loss of knowledge-base functionality); and add input validation that rejects absolute paths and normalizes submitted filenames against the configured storage root before any file operation. Complement these with least-privilege filesystem ACLs, file-integrity monitoring or audit rules alerting on writes outside the Langflow workspace, and verified backups so overwritten files can be restored.
Affected VersionsENISA
Affects your stack
Find out whether the stacks and servers you run are on an affected version.
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Langflow is a Python-based low-code platform for building LLM applications, exposing a FastAPI-style backend whose knowledge-base module (src/backend/base/langflow/api/v1/knowledge_bases.py) handles creation of knowledge bases over HTTP. The defect is CWE-22 (Improper Limitation of a Pathname to a Restricted Directory / Path Traversal), specifically the absolute-path variant: rather than requiring ../ sequencing, the handler accepts a path or filename that can be absolute and passes it directly into filesystem creation routines, so no directory boundary is enforced and any path writable by the Langflow process becomes a valid target. Because the injected value controls the destination rather than the content rendering, the modus operandi is arbitrary file write/overwrite (integrity, I:H) rather than data disclosure (C:N), and the independent assessment's A:L reflects the potential to disrupt the service by clobbering files it depends on, while the vendor vector lists A:N. The provided CPE string is unpopulated (cpe:2.3:a:n/a:n/a:*:*:*:*:*:*:*:*) and EUVD records the affected version as n/a, so version attribution rests on the CVE description's 'up to 1.8.4' statement and the vulnerable code path at line 51 rather than on enumerative product metadata.
Know when Langflow gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:N
Network HTTP endpoint with low complexity; PR:N retained per the public-facing description though deployments may need auth; arbitrary file write gives I:H and plausible A:L via overwriting critical files; no data disclosure so C:N.
Primary rating from Vendor (mitre).
CVSS VectorVendor: mitre
Lifecycle Timeline
3DescriptionCVE.org
langflow-ai langflow v1.8.4 is affected by: Directory Traversal. The impact is: Arbitrary file write outside the intended workspace or storage boundary.. The component is: src/backend/base/langflow/api/v1/knowledge_bases.py:knowledge_bases-create_knowledge_base-a-live-http-post-to-create-knowledge-base. The attack vector is: Attack surface: HTTP or browser-backed service path. A public-facing upload or HTTP route handler forwards an attacker-controlled path or filename into host file creation without any visible boundary enforcement. ¶¶ A weakness has been identified in langflow-ai langflow up to 1.8.4. langflow contains an absolute path traversal vulnerability in knowledge_bases-create_knowledge_base-a-live-http-post-to-create-knowledge-base (src/backend/base/langflow/api/v1/knowledge_bases.py:51). An attacker can write or overwrite files outside the intended working directory by providing absolute paths in the knowledge base creation endpoint.
Executive ViewAI
Langflow through v1.8.4 contains a high-severity absolute path traversal flaw in its knowledge-base creation endpoint that allows a remote actor to write or overwrite files reachable by the Langflow service account. The CVSS vector rates the issue as unauthenticated (PR:N), although deployments that enforce API login may require a valid session in practice, and no public exploit has been identified at time of analysis. Because no vendor-released patch is available, affected organizations face potential data corruption, unauthorized configuration changes, or further compromise depending on the service account's operating-system privileges, making exposure reduction and monitoring a leadership priority.
Restrict access to POST /api/v1/knowledge_bases to trusted networks and authenticated users, enforce least privilege for the Langflow service account, deploy WAF/reverse-proxy rules blocking absolute path and traversal payloads, enable file integrity monitoring on reachable paths, and maintain offline backups.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Path traversal exploits occur when applications use user-controlled input to construct file system paths without proper validation.
Langflow before 1.3.0 allows unauthenticated remote code injection through the /api/v1/validate/code endpoint, enabling
Langflow has a third RCE vulnerability via exec_globals (EPSS 10.0%) allowing inclusion of untrusted code that executes
Account takeover and remote code execution in Langflow AI workflow platform versions ≤1.6.9 due to a CORS misconfigurati
Langflow has a code injection vulnerability in the code component (EPSS 2.6%) enabling remote code execution through the
Langflow has an eval injection in eval_custom_component_code (EPSS 2.0%) enabling remote code execution through crafted
Code injection in Langflow CSV Agent node before 1.8.0. The node hardcodes allow_dangerous_code=True, enabling arbitrary
Langflow through 0.6.19 allows remote code execution if untrusted users are able to reach the "POST /api/v1/custom_compo
OS command injection in Langflow's MCP stdio transport allows an attacker who can reach the MCP server configuration end
langflow <=1.0.18 is vulnerable to Remote Code Execution (RCE) as any component provided the code functionality and the
langflow v1.0.12 was discovered to contain a remote code execution (RCE) vulnerability via the PythonCodeTool component.
Langflow before 1.7.0.dev45 exposes multiple API endpoints without authentication, allowing unauthenticated access to us
An unauthenticated shell injection vulnerability exists in Langflow's GitHub Actions CI/CD workflows, allowing attackers
EUVD-2026-91119
GHSA-5rwq-2vwh-7g73
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB