PT-2026-51212

Name of the Vulnerable Software and Affected Versions BerriAI litellm versions prior to 1.82.3 Description A server-side request forgery SSRF exists in the MCP OpenAPI Spec Loader component. The issue occurs within the load openapi spec async function located in the litellm/proxy/ experimental/mc...

MINI-Q949-MF84-3MCP

Bulletin has no description...

flowise-mcp-env-case-bypass-poc

Flowise 3.1.2 Custom MCP Environment Variable Case Bypass PoC...

GHSA-V3F4-W7R7-V3HM Uni-CLI: Legacy HTTP MCP transport accepted browser-originated localhost requests

Impact Uni-CLI versions before 0.225.2 exposed the legacy JSON-RPC-over-HTTP MCP transport on loopback without validating browser Origin headers before routing requests. A malicious web page could send a CORS simple POST request, such as text/plain, to the local /mcp endpoint and deliver a JSON-R...

Origin Validation Error

Overview dbt-mcp is an A MCP Model Context Protocol server for interacting with dbt resources. Affected versions of this package are vulnerable to Origin Validation Error in the GET /dbtplatformcontext endpoint, which is exposed without authentication or host-origin validation. An attacker can...

CVE-2026-48787

gin-vue-admin is an AI-assisted basic development platform. In version 2.9.1, an authenticated attacker with access to the code-generation feature and MCP management interface can exploit this vulnerability by injecting attacker-controlled Go source code through POST /autoCode/addFunc, and then...

CVE-2026-48787

CVE-2026-48787 affects gin-vue-admin (AI-assisted basic development platform) in version 2.9.1. An authenticated attacker with access to the code-generation feature and MCP management interface can inject attacker-controlled Go source code via POST /autoCode/addFunc, then trigger a rebuild of the...

CVE-2026-48787 gin-vue-admin vulnerable to RCE

gin-vue-admin is an AI-assisted basic development platform. In version 2.9.1, an authenticated attacker with access to the code-generation feature and MCP management interface can exploit this vulnerability by injecting attacker-controlled Go source code through POST /autoCode/addFunc, and then...

CVE-2026-48787

gin-vue-admin is an AI-assisted basic development platform. In version 2.9.1, an authenticated attacker with access to the code-generation feature and MCP management interface can exploit this vulnerability by injecting attacker-controlled Go source code through POST /autoCode/addFunc, and then...

CVE-2026-48787 gin-vue-admin vulnerable to RCE

gin-vue-admin is an AI-assisted basic development platform. In version 2.9.1, an authenticated attacker with access to the code-generation feature and MCP management interface can exploit this vulnerability by injecting attacker-controlled Go source code through POST /autoCode/addFunc, and then...

CVE-2026-48774

Summary : ProxySQL 3.0.0–3.0.8 allows read-only requests to execute multi-statement backends, enabling unintended writes via the MCP run_sql_readonly tool. The input validator uses a blacklist/allowlist on the first statement, but then runs the full string against a backend connection created wit...

CVE-2026-48774 ProxySQL MCP run_sql_readonly executes side-effecting MySQL multi-statements despite read-only contract

ProxySQL is a proxy for MySQL and its forks, as well as PostgreSQL. In versions 3.0.0 through 3.0.8, ProxySQL's GenAI/MCP runsqlreadonly tool violates its documented read-only contract for MySQL targets. The tool validates only the full input string with a substring blacklist and first-keyword...

CVE-2026-49291

mcp-memory-service is a semantic memory layer for AI applications. Prior to version 10.65.3, the HTTP MCP JSON-RPC endpoint at /mcp requires only OAuth read scope for all requests, then dispatches tools/call directly to handlers that include mutating tools. A read-only OAuth client can call...

CVE-2026-49291

mcp-memory-service (semantic memory layer for AI apps) exposed the HTTP MCP JSON-RPC endpoint at /mcp such that OAuth read scope allowed mutating actions. Before patch 10.65.3, a read-only OAuth client could invoke tools/call to reach store_memory and delete_memory, bypassing REST write scope che...

CVE-2026-49291 mcp-memory-service: OAuth read-only clients can write and delete memories through MCP tools/call

mcp-memory-service is a semantic memory layer for AI applications. Prior to version 10.65.3, the HTTP MCP JSON-RPC endpoint at /mcp requires only OAuth read scope for all requests, then dispatches tools/call directly to handlers that include mutating tools. A read-only OAuth client can call...

CVE-2026-49291

mcp-memory-service is a semantic memory layer for AI applications. Prior to version 10.65.3, the HTTP MCP JSON-RPC endpoint at /mcp requires only OAuth read scope for all requests, then dispatches tools/call directly to handlers that include mutating tools. A read-only OAuth client can call...

Weekly Metasploit Update: NTLM Relay Priv Esc, MCP Server Integration, Paperclip AI RCE Chain, and more

This week's release includes five new modules, including a full unauthenticated RCE chain for Paperclip AI and a VS Code extension persistence technique. On the post-exploitation side, the new windows/local/ntlmrelay2self module coerces the local machine account to authenticate via...

MINI-4MCP-M74H-7CW4

Bulletin has no description...

GHSA-JV2H-4P9V-WF5W ouroboros-ai: Incomplete fix of CVE-2026-47211: untrusted project .env can still reach RCE via omitted execution-routing keys

Impact The CVE-2026-47211 fix 0.39.0 added UNTRUSTEDENVDENYLIST to stop an untrusted project-directory .env from redirecting execution. The denylist was incomplete — several execution-routing keys of the same RCE class were omitted, so a malicious cloned repo can still reach arbitrary command...

CVE-2026-49357

Line Desktop MCP is a project that, while unaffiliated with the official line-bot-mcp-server, allows users to directly operate the LINE Desktop application on Windows or Mac via MCP. line-desktop-mcp supports a --http-mode Streamable HTTP transport for use with clients such as n8n. In this mode t...