Описание
BerriAI LiteLLM before 1.101.0-rc.1 contains a tenant isolation bypass vulnerability in the semantic cache layer that allows authenticated users to read other tenants' cached responses by exploiting a metadata key mismatch between _get_semantic_cache_tenant_scope() and _get_metadata_variable_name(). Attackers holding a valid virtual key can submit semantically similar prompts on affected routes such as /v1/responses and /bedrock/* to retrieve cached responses containing other tenants' personally identifiable information, financial data, or source code, and can cause agentic front-ends to auto-execute attacker-supplied tool calls under victim credentials by returning cached function_call or tool_calls payloads to a different principal.
A flaw was found in LiteLLM. Due to a metadata key mismatch in the semantic cache component, an authenticated remote attacker can bypass tenant isolation by submitting semantically similar queries. This issue leads to information disclosure, allowing unauthorized access to cached responses containing other tenants' sensitive data, and may cause connected automated agents to execute unauthorized actions under victim credentials.
Отчет
This vulnerability is rated as Important severity because authenticated multi-tenant users can bypass isolation boundaries to access confidential response data and trigger unauthorized tool execution across distinct accounts. While exploitation requires valid API access to affected proxy routes, the shared caching mechanism compromises confidentiality and integrity across AI workloads. Deployments serving single-tenant workloads or environments where semantic caching is not shared between untrusted tenants face substantially reduced exposure.
Меры по смягчению последствий
Mitigation for this issue is either not available or the currently available options do not meet the Red Hat Product Security criteria comprising ease of use and deployment, applicability to widespread installation base, or stability.
Затронутые пакеты
| Платформа | Пакет | Состояние | Рекомендация | Релиз |
|---|---|---|---|---|
| Exploit Intelligence | exploit-intelligence/vulnerability-analysis-rhel9 | Affected | ||
| Lightspeed Core | lightspeed-core/lightspeed-stack-rhel9 | Affected | ||
| Red Hat Ansible Automation Platform 2 | ansible-automation-platform-26/lightspeed-chatbot-rhel9 | Affected | ||
| Red Hat Ansible Automation Platform 2 | ansible-automation-platform-27/lightspeed-chatbot-rhel9 | Affected | ||
| Red Hat OpenShift AI (RHOAI) | rhoai/odh-llama-stack-core-rhel9 | Affected | ||
| Red Hat OpenShift AI 3.4 | rhoai/odh-trustyai-garak-lls-provider-dsp-rhel9 | Fixed | RHSA-2026:84734 | 19.10.2026 |
| Red Hat OpenShift AI 3.5 | rhoai/odh-mlflow-rhel9 | Fixed | RHSA-2026:95872 | 06.11.2026 |
| Red Hat OpenShift AI 3.5 | rhoai/odh-th-torch-cpu-py312-rhel9 | Fixed | RHSA-2026:95872 | 06.11.2026 |
| Red Hat OpenShift AI 3.5 | rhoai/odh-th-torch-cuda-py312-rhel9 | Fixed | RHSA-2026:95872 | 06.11.2026 |
Показывать по
Ссылки на источники
Дополнительная информация
Статус:
EPSS
7.7 High
CVSS3
Связанные уязвимости
BerriAI LiteLLM before 1.101.0-rc.1 contains a tenant isolation bypass vulnerability in the semantic cache layer that allows authenticated users to read other tenants' cached responses by exploiting a metadata key mismatch between _get_semantic_cache_tenant_scope() and _get_metadata_variable_name(). Attackers holding a valid virtual key can submit semantically similar prompts on affected routes such as /v1/responses and /bedrock/* to retrieve cached responses containing other tenants' personally identifiable information, financial data, or source code, and can cause agentic front-ends to auto-execute attacker-supplied tool calls under victim credentials by returning cached function_call or tool_calls payloads to a different principal.
BerriAI LiteLLM before 1.101.0-rc.1 contains a tenant isolation bypass vulnerability in the semantic cache layer that allows authenticated users to read other tenants' cached responses by exploiting a metadata key mismatch between _get_semantic_cache_tenant_scope() and _get_metadata_variable_name(). Attackers holding a valid virtual key can submit semantically similar prompts on affected routes such as /v1/responses and /bedrock/* to retrieve cached responses containing other tenants' personally identifiable information, financial data, or source code, and can cause agentic front-ends to auto-execute attacker-supplied tool calls under victim credentials by returning cached function_call or tool_calls payloads to a different principal.
EPSS
7.7 High
CVSS3