CVE-2026-89032 BerriAI LiteLLM < 1.101.0-rc.1 Tenant Isolation Bypass via Semantic Cache Layer

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 getsemanticcachetenantscope and getmetadatavariablename. Attackers...

CVE-2026-64651: Incorrect Authorization

The @ai-sdk/harness-opencode tool connects HarnessAgent to OpenCode through a sandboxed bridge. Prior to version 1.0.28, the tool relay authorizes requests from any process whose command line contains an allowed helper script path host-tool-mcp.mjs. This allows untrusted code executing in the...

Authorization Bypass Through User-Controlled Key

Overview mcp is a The official Ruby SDK for Model Context Protocol servers and clients Affected versions of this package are vulnerable to Authorization Bypass Through User-Controlled Key in lib/mcp/server/transports/streamablehttptransport.rb, which processes tool execution requests to...

Neo4j MCP Clients & Servers 安全漏洞

Neo4j MCP Clients & Servers is a protocol for managing large language model contexts in the Neo4j Contrib open source. A security vulnerability exists in Neo4j MCP Clients & Servers, which stems from a DNS rebinding vulnerability that could lead to bypassing same-origin policy protections and...

LLMail-Inject: a Dataset from a Realistic Adaptive Prompt Injection Challenge

Indirect Prompt Injection attacks exploit the inherent limitation of Large Language Models LLMs to distinguish between instructions and data in their inputs. Despite numerous defense proposals, the systematic evaluation against adaptive adversaries remains limited, even when successful attacks ca...