CVE-2023-46229: LangChain: SSRF in URL loader exposes internal network

HIGHACTIVELY EXPLOITEDPoC AVAILABLE

Published October 19, 2023

CISO Take

LangChain's RecursiveURLLoader follows redirects from external to internal addresses, enabling attackers to reach cloud metadata services (AWS IMDS at 169.254.169.254), internal APIs, and credential stores. Any RAG pipeline or agent deployment that processes user-supplied or untrusted URLs via document loaders is exposed. Upgrade to LangChain 0.0.317+ immediately and enforce network-level blocks on RFC 1918 and link-local ranges for LangChain hosts.

What is the risk?

High risk (CVSS 8.8). Attack complexity is low and no privileges are required — an attacker only needs to supply a URL that triggers a redirect to an internal target. The 'UI:Required' reflects that a user or automated pipeline must invoke the loader, which in practice is the default operating mode for RAG ingestion jobs and AI agents. The primary escalation path is cloud credential theft via metadata services, which can result in full cloud account takeover. LangChain's ubiquity across AI/ML deployments significantly amplifies blast radius.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
LangChain	pip	—	No patch
148.8KOpenSSF 5.73.2K dependentsPushed 5d ago30% patched~164d to patchFull package profile →

Do you use LangChain? You're affected.

How severe is it?

CVSS 3.1

8.8 / 10

EPSS

47.6%

chance of exploitation in 30 days

Higher than 99% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Actively Exploited

Sophistication

Trivial

Exploitation Confidence

high

✓CISA KEV (active exploitation confirmed)

○Public PoC indexed (trickest/cve)

○EPSS exploit prediction: 48%

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AVNetwork

ACLow

PRNone

UIRequired

SUnchanged

CHigh

IHigh

AHigh

What should I do?

5 steps

PATCH
Upgrade LangChain to >= 0.0.317 immediately (commit 9ecb724 addresses this).
NETWORK CONTROL
Block outbound HTTP/S from LangChain processes to RFC 1918 (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16) and link-local (169.254.0.0/16) ranges at the firewall or security group level.
INPUT VALIDATION
If patching is delayed, implement a domain allowlist for all inputs to RecursiveURLLoader — reject any URL not on the allowlist before instantiating the loader.
CLOUD HARDENING
Enable IMDSv2 (AWS), disable legacy IMDS, and apply equivalent protections on GCP/Azure to limit credential exposure via SSRF.
DETECTION
Alert on HTTP requests from LangChain processes to private IP ranges, metadata service IPs, or internal hostnames; review application logs for unexpected content in loader outputs.

What does CISA's SSVC say?

DecisionTrack

Exploitationnone

AutomatableNo

Technical Impacttotal

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Data ExtractionAuth BypassFrameworkRAGAML.T0010.001- AI Software AML.T0025- Exfiltration via Cyber Means AML.T0037- Data from Local System AML.T0049- Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

EU AI Act

Art. 15-Accuracy, Robustness and Cybersecurity

ISO 42001

A.9.3-AI System Security

NIST AI RMF

MANAGE-2.2-Manage AI Risk Responses

OWASP LLM Top 10

LLM05-Supply Chain VulnerabilitiesLLM07-Insecure Plugin Design

Frequently Asked Questions

What is CVE-2023-46229?

Is CVE-2023-46229 actively exploited?

Yes, CVE-2023-46229 is confirmed actively exploited and listed in CISA Known Exploited Vulnerabilities catalog.

How to fix CVE-2023-46229?

1. PATCH: Upgrade LangChain to >= 0.0.317 immediately (commit 9ecb724 addresses this). 2. NETWORK CONTROL: Block outbound HTTP/S from LangChain processes to RFC 1918 (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16) and link-local (169.254.0.0/16) ranges at the firewall or security group level. 3. INPUT VALIDATION: If patching is delayed, implement a domain allowlist for all inputs to RecursiveURLLoader — reject any URL not on the allowlist before instantiating the loader. 4. CLOUD HARDENING: Enable IMDSv2 (AWS), disable legacy IMDS, and apply equivalent protections on GCP/Azure to limit credential exposure via SSRF. 5. DETECTION: Alert on HTTP requests from LangChain processes to private IP ranges, metadata service IPs, or internal hostnames; review application logs for unexpected content in loader outputs.

What systems are affected by CVE-2023-46229?

This vulnerability affects the following AI/ML architecture patterns: RAG pipelines, agent frameworks, document processing pipelines, LLM application frameworks, knowledge base ingestion pipelines.

What is the CVSS score for CVE-2023-46229?

CVE-2023-46229 has a CVSS v3.1 base score of 8.8 (HIGH). The EPSS exploitation probability is 47.63%.

What is the AI security impact?

Affected AI Architectures

RAG pipelinesagent frameworksdocument processing pipelinesLLM application frameworksknowledge base ingestion pipelines

MITRE ATLAS Techniques

AML.T0010.001AI Software

AML.T0025Exfiltration via Cyber Means

AML.T0037Data from Local System

AML.T0049Exploit Public-Facing Application

Compliance Controls Affected

EU AI Act: Art. 15

ISO 42001: A.9.3

NIST AI RMF: MANAGE-2.2

OWASP LLM Top 10: LLM05, LLM07

What are the technical details?

Original Advisory

LangChain before 0.0.317 allows SSRF via document_loaders/recursive_url_loader.py because crawling can proceed from an external server to an internal server.

Exploitation Scenario

An attacker controls an external web server at attacker-site.com. They submit this URL as a document source to a LangChain-based RAG pipeline that indexes web content for a corporate AI assistant. When RecursiveURLLoader crawls attacker-site.com, the server responds with HTTP 301 redirecting to http://169.254.169.254/latest/meta-data/iam/security-credentials/MyRole. LangChain follows the redirect and fetches the AWS IAM temporary credentials. The attacker extracts these from the indexed content, application logs, or error messages, then uses them to access the organization's S3 buckets containing training data, model weights, and proprietary documents — or to pivot into the broader AWS environment.

Weaknesses (CWE)

CWE-918Server-Side Request ForgeryPrimary

CWE-918 — Server-Side Request Forgery: The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination.

[Implementation, Operation] If there is a known list of applications, IP, hostname, and/or domains to which the product is allowed to communicate, then use allowlists to ensure that outgoing requests are restricted to these known entities. The allowlists could be in the application code or possibly enforced by firewalls or network segmentation [REF-1511].
[Implementation, Operation] If the product is expected to be able to connect to arbitrary IP addresses or domains, consider using a denylist that includes private network addresses, internal domain names, etc. The denylist could be in the application code or possibly enforced by firewalls or network segmentation [REF-1511].

Source: MITRE CWE corpus.