CVE-2026-51888: Langflow: absolute path traversal allows arbitrary file write
HIGHPoC AVAILABLECISA: TRACK*Langflow 1.8.4 and earlier contain an absolute path traversal flaw in the knowledge base creation endpoint (knowledge_bases.py), which lets an unauthenticated remote attacker write or overwrite files anywhere the Langflow process has permissions. It is rated High (CVSS 7.5, network, low complexity, no privileges, no user interaction), but the integrity-only impact is limited on paper: arbitrary file overwrite in a service that often runs with broad filesystem access can realistically be chained into code execution. Exploitation likelihood is currently low-to-moderate (EPSS 0.56%, around the 55th percentile), the CVE is not in CISA KEV, and there is no public exploit or Nuclei template. The only reference is a researcher gist, so details are public enough to be weaponized quickly. Langflow has a history of being targeted, so treat it as a patch-soon item. Upgrade past 1.8.4 as soon as a fixed release is available (confirm with the vendor advisory, since no fix version is listed in the data), and until then do not expose Langflow directly to the internet or untrusted networks. Also restrict access to the knowledge base endpoint at the reverse proxy.
What is the risk?
Medium-High. Exploitability is excellent (network, low complexity, no auth, no interaction), but impact is limited to integrity (I:H, C:N, A:N) and there is no evidence of exploitation in the wild (not in KEV, EPSS 0.56%, no public exploit or Nuclei template). The practical risk rises sharply for internet-exposed Langflow instances running as root or in containers with writable application, cron or SSH paths, where file overwrite can become remote code execution. No CWE or affected-version metadata is formally listed, so confidence in the exact scope is medium.
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 Langflow to the first release after 1.8.4 that fixes this issue (verify in the vendor GitHub advisory/releases). 2) Until patched, remove public exposure: place Langflow behind a VPN, authenticated reverse proxy or IP allowlist, and block or restrict POST requests to the knowledge base creation endpoint (/api/v1/knowledge_bases). 3) Run the service as a non-root user in a container with a read-only root filesystem and writable mounts limited to the data directory. 4) Detection: alert on requests to the knowledge base creation endpoint containing absolute paths (leading '/' or drive letters) or '..' sequences, on unexpected file modifications outside the Langflow data directory (file integrity monitoring on app code, cron, SSH authorized_keys), and on new files owned by the Langflow user. 5) Audit instances for unexpected changes and rotate secrets if tampering is suspected.
What does CISA's SSVC say?
Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-51888?
Langflow 1.8.4 and earlier contain an absolute path traversal flaw in the knowledge base creation endpoint (knowledge_bases.py), which lets an unauthenticated remote attacker write or overwrite files anywhere the Langflow process has permissions. It is rated High (CVSS 7.5, network, low complexity, no privileges, no user interaction), but the integrity-only impact is limited on paper: arbitrary file overwrite in a service that often runs with broad filesystem access can realistically be chained into code execution. Exploitation likelihood is currently low-to-moderate (EPSS 0.56%, around the 55th percentile), the CVE is not in CISA KEV, and there is no public exploit or Nuclei template. The only reference is a researcher gist, so details are public enough to be weaponized quickly. Langflow has a history of being targeted, so treat it as a patch-soon item. Upgrade past 1.8.4 as soon as a fixed release is available (confirm with the vendor advisory, since no fix version is listed in the data), and until then do not expose Langflow directly to the internet or untrusted networks. Also restrict access to the knowledge base endpoint at the reverse proxy.
Is CVE-2026-51888 actively exploited?
Proof-of-concept exploit code is publicly available for CVE-2026-51888, increasing the risk of exploitation.
How to fix CVE-2026-51888?
1) Upgrade Langflow to the first release after 1.8.4 that fixes this issue (verify in the vendor GitHub advisory/releases). 2) Until patched, remove public exposure: place Langflow behind a VPN, authenticated reverse proxy or IP allowlist, and block or restrict POST requests to the knowledge base creation endpoint (/api/v1/knowledge_bases). 3) Run the service as a non-root user in a container with a read-only root filesystem and writable mounts limited to the data directory. 4) Detection: alert on requests to the knowledge base creation endpoint containing absolute paths (leading '/' or drive letters) or '..' sequences, on unexpected file modifications outside the Langflow data directory (file integrity monitoring on app code, cron, SSH authorized_keys), and on new files owned by the Langflow user. 5) Audit instances for unexpected changes and rotate secrets if tampering is suspected.
What systems are affected by CVE-2026-51888?
This vulnerability affects the following AI/ML architecture patterns: LLM orchestration frameworks, RAG pipelines, agent frameworks, knowledge base / vector ingestion pipelines, low-code AI workflow platforms.
What is the CVSS score for CVE-2026-51888?
CVE-2026-51888 has a CVSS v3.1 base score of 7.5 (HIGH). The EPSS exploitation probability is 0.64%.
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0010.001AI SoftwareAML.T0049Exploit Public-Facing ApplicationAML.T0112Machine CompromiseCompliance Controls Affected
What are the technical details?
Original Advisory
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.
Exploitation Scenario
An attacker scans for internet-exposed Langflow instances and sends an unauthenticated POST to the knowledge base creation endpoint, supplying an absolute path such as a Python module in the site-packages directory, a cron file or an SSH authorized_keys file as the knowledge base name or location. The server writes attacker-influenced content to that path. The attacker overwrites a file that is executed on the next restart or schedule, gaining code execution in the AI workflow host. From there they read API keys for LLM providers and vector databases from the environment, and manipulate flows and RAG data.
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:NTimeline
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