VYPR
High severity8.6NVD Advisory· Published Jun 15, 2026

CVE-2026-47835

CVE-2026-47835

Description

Spring AI Vector Stores fail to sanitize special characters in metadata filtering, enabling arbitrary query execution in Elasticsearch, OpenSearch, and GemFire VectorDB.

AI Insight

LLM-synthesized narrative grounded in this CVE's description and references.

Spring AI Vector Stores fail to sanitize special characters in metadata filtering, enabling arbitrary query execution in Elasticsearch, OpenSearch, and GemFire VectorDB.

Vulnerability

Spring AI Vector Stores in versions 1.0.0 through 1.0.x and 1.1.0 through 1.1.x do not properly sanitize special characters in metadata filtering. This flaw affects the spring-ai-elasticsearch-store, spring-ai-opensearch-store, and spring-ai-gemfire-store components, allowing an attacker to inject arbitrary query syntax into the underlying vector database queries [1].

Exploitation

An attacker with network access to the vector store endpoint can craft a malicious metadata filter containing special characters (e.g., quotes, operators) that bypasses intended filtering logic. No authentication or user interaction is required, as the vulnerability is triggered purely by sending a crafted request to the affected service [1].

Impact

Successful exploitation leads to unauthorized execution of arbitrary queries against the Elasticsearch, OpenSearch, or GemFire VectorDB backends. This can result in high confidentiality impact (reading sensitive vector data) and limited integrity/availability impact (modifying or deleting data). The CVSS v3.1 score is 8.6 (High) [1].

Mitigation

Users should upgrade to the fixed versions: Spring AI 1.0.9 for the 1.0.x line and Spring AI 1.1.8 for the 1.1.x line. No other mitigation steps are necessary, and no workarounds are provided [1].

AI Insight generated on Jun 15, 2026. Synthesized from this CVE's description and the cited reference URLs; citations are validated against the source bundle.

Affected products

3

Patches

0

No patches discovered yet.

Vulnerability mechanics

No source-code context for this CVE — mechanics is only generated when we can read the actual fix diff. Without that, the four sections (root cause, attack vector, affected code, fix) would be speculation rather than analysis.

References

1

News mentions

0

No linked articles in our index yet.