Vulnerability in Langroid Framework's Neo4jChatAgent Allows Prompt Injection
CVE-2026-55615

9.2CRITICAL

Key Information:

Vendor

Langroid

Status
Langroid
Vendor
CVE Published:
9 July 2026

What is CVE-2026-55615?

The Langroid framework, specifically its Neo4jChatAgent, contains a vulnerability that allows prompt injection due to the lack of validation when processing LLM-generated Cypher queries. This issue enables attackers to manipulate the query text, potentially leading to the unauthorized viewing or destruction of graph data. Furthermore, if APOC or dbms.security procedures are enabled, attackers can gain access to OS commands and filesystem operations. The vulnerabilities were not addressed in earlier versions, including a previous similar issue fixed in version 0.63.0. A solution is provided in version 0.65.5, which includes necessary validations.

Affected Version(s)

langroid < 0.65.5

References

https://github.com/langroid/langroid/security/advisories/...
x_refsource_CONFIRM
https://github.com/langroid/langroid/commit/5a3097d9dd637...
x_refsource_MISC

CVSS V4

Score:
9.2
Severity:
CRITICAL
Confidentiality:
High
Integrity:
High
Availability:
High
Attack Vector:
Network
Attack Complexity:
Low
Attack Required:
Physical
Privileges Required:
Undefined
User Interaction:
None

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.