CVE-2026-109440

Deferred Deferred - Pending Action

Indirect Prompt Injection in LangAlpha

Publication date: 2026-10-20

Last updated on: 2026-10-22

Assigner: VulnCheck

Description

LangAlpha through 2026.09.07 contains an indirect prompt injection vulnerability that allows remote attackers to execute shell commands or Python by planting instructions in web content retrieved by the agent. Because Bash and ExecuteCode lack an allow-list and human-in-the-loop approval, attackers can run code in the workspace sandbox and exfiltrate vault secrets from _internal/.vault_secrets.json.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published

2026-10-20

Last Modified

2026-10-22

Generated

2026-11-09

AI Q&A

2026-10-20

EPSS Evaluated

2026-11-07

NVD

CVE-2026-109440

EUVD

EUVD-2026-104680

Affected Vendors & Products

Vendor	Product	Version / Range
ginlix-ai	LangAlpha	0

Helpful Resources

Exploitability

CWE

KEV

CWE ID	Description
CWE-UNKNOWN

Attack-Flow Graph

Executive Summary

LangAlpha through 2026.09.07 has an indirect prompt injection vulnerability. Attackers can embed malicious instructions in web content retrieved by the agent. Because Bash and ExecuteCode lack an allow-list and human approval, they can execute shell commands or Python code in the workspace sandbox. This allows exfiltration of vault secrets from _internal/.vault_secrets.json.

Detection Guidance

Check for unauthorized shell commands or Python code execution in logs from LangAlpha agents. Inspect network traffic for suspicious outbound connections from the sandbox environment. Review files modified by the agent, especially _internal/.vault_secrets.json for unexpected access.

Impact Analysis

An attacker could run arbitrary code in your workspace, steal sensitive secrets stored in vault files, or perform unauthorized actions. The lack of input validation and approval means any web content processed by the agent could be malicious.

Compliance Impact

This vulnerability could lead to unauthorized access to sensitive data, violating GDPR (data protection) and HIPAA (health information privacy). Exfiltration of secrets like credentials or personal data risks non-compliance with these regulations.

Mitigation Strategies

Upgrade LangAlpha to a version beyond 2026.09.07. Implement allow-lists for Bash and ExecuteCode tools. Add human-in-the-loop approval for code execution. Restrict sandbox permissions to prevent access to _internal/.vault_secrets.json.

Hi! I’m here to help you understand CVE-2026-109440. Ask me anything about the vulnerability, its impact, or mitigation strategies.

0/70

EPSS Chart

Security Radar

News mentioning this vulnerability

Loading related security news...