LLM-MCP-Security-Field-Guide

🛡️ Guida alla Sicurezza AI — Sicurezza LLM + MCP Il riferimento di sicurezza più completo, aggiornato e pratico per applicazioni LLM e implementazioni del Model Context Protocol MCP. Copre CVE reali, modelli di attacco attuali, framework OWASP, strumenti di red teaming e checklist attuabili —...

agent-scan

Snyk Agent Scan Entdecken und scannen Sie Agent-Komponenten auf Ihrem Rechner auf Prompt-Injections und Schwachstellen einschließlich Agenten, MCP-Server und Skills. Hinweis: Die CLI-Ausgabe ist experimentell und kann sich ändern Agent Scan v0.5.x zur Einstellung vorgesehen Die Rohausgabe dieser...

mcp-server-attestation

Parte delloStudioMeyer MCP Stack — Costruito a Maiorca 🌴 · ⭐ se lo usi mcp-server-attestation Rafforzamento della supply chain di Layer-2 per i server Model Context Protocol. Manifest degli strumenti firmati con Ed25519, attestazione runtime degli spawn, sanitizer degli argomenti con negazione...

EUVD-2026-67022

The mcp-router CLI served its MCP aggregator on every interface and enforced authentication only when the operator asked for it. The serve command in apps/cli/src/commands/serve.ts defaulted its host to the all-interfaces address on a fixed port, and required a token only when the corresponding...

pkgxray

pkgxray — sicurezza pre-installazione per pacchetti npm, server MCP e agenti AI Ispeziona un pacchetto npm o un server MCP prima di installarlo o di connetterti ad esso, e ottieni un verdetto deterministico e basato su prove: SAFE, REVIEW, o BLOCK. Analisi statica locale, a zero dipendenze — le...

CVE

Tabella delle CVE Numero CVE| Tipo| Azienda/Prodotto| Descrizione| Link articolo| Link repository ---|---|---|---|---|--- CVE-2026-9680| SQLi| Alibaba RDS MCP Servers| TBA| TBA| TBA CVE-2026-50335| LPE| Windows July 2026| TBA| TBA| TBA GHSA-73cv-556c-w3g6| SQLi| Apache Pinot MCP 3.1.0| La mancata...

rbinmcp

rbinmcp - Ritirato Questo repository è stato ritirato e suddiviso in server MCP separati e mantenibili: r2 backend - github.com/kirkderp/rbinr2 Ghidra backend - github.com/kirkderp/rbinghidra ILSpy .NET backend - github.com/kirkderp/rbinilspy Tutto il codice e le crate sono stati rimossi per...

aimap

AIMap 面向暴露的AI代理基础设施的互联网规模发现与安全测试平台。 AIMap 可发现、指纹识别并安全测试公开暴露的 AI 端点——包括 MCP 服务器、Ollama 实例、vLLM/LiteLLM 代理、LangServe 链、Gradio 应用、ComfyUI 节点等。可以将其视为 Shodan，但专门为 AI 代理攻击面打造。 由 Bishop Fox 构建。 警告 此工具仅用于授权的渗透测试和安全研究 。您只能对您拥有或已获得明确书面许可的系统使用 AIMap。未经授权访问计算机系统是非法的。Bishop Fox...

mcp-xray

MCP X-Ray Übersicht MCP X-Ray ist eine einheitliche Open-Source-Sicherheitsscan- und Penetrationstestlösung für Model Context Protocol MCP-Server. Es erzeugt produktionsreife SARIF-Berichte für die nahtlose Integration in Sicherheitstools und CI/CD-Pipelines. Suchergebnisse können optional auf...

vulnerable-mcp-servers-lab

Laboratorio di server MCP vulnerabili Questo repository contiene implementazioni volutamente vulnerabili di server Model Context Protocol MCP sia locali che remoti. Ogni server risiede nella propria cartella e include un README.md dedicato con tutti i dettagli su cosa fa , come eseguirlo e come...

agent-opfor

Emulazione dell'avversario open-source per agenti AI, app LLM e server MCP. Testa la tua IA come farebbe un vero attaccante — dal tuo CLI, dal tuo IDE o da un'estensione del browser che chiunque nel tuo team può usare. Sito web · Documentazione · GitHub · ·...

CVE-2026-71424

CVE-2026-71424 is a critical cross-user OAuth token leak in Onyx , an open-source AI platform. The GET /api/mcp/servers and GET /api/mcp/servers/persona/{persona_id} endpoints expose another user's OAuth Authorization header. The root cause is that OnyxTokenStorage.set_tokens and OnyxTokenStorage...

CVE-2026-71424 Onyx: Cross-user OAuth-token leak via /api/mcp/servers* for per-user MCP servers

Onyx is an open-source AI platform. Prior to 3.1.10, 3.2.14, and 4.0.0, Onyx's GET /api/mcp/servers and GET /api/mcp/servers/persona/personaid endpoints expose another user's OAuth Authorization header because OnyxTokenStorage.settokens and OnyxTokenStorage.setclientinfo in...

CVE-2026-19770

A vulnerability was identified in feedmob fm-mcp-servers 0.0.3. Affected by this vulnerability is the function downloadReport of the file src/smadex-reporting/src/index.ts of the component Download Endpoint. The manipulation of the argument downloadUrl leads to server-side request forgery. The...

CVE-2026-19770 feedmob fm-mcp-servers Download Endpoint index.ts downloadReport server-side request forgery

A vulnerability was identified in feedmob fm-mcp-servers 0.0.3. Affected by this vulnerability is the function downloadReport of the file src/smadex-reporting/src/index.ts of the component Download Endpoint. The manipulation of the argument downloadUrl leads to server-side request forgery. The...

CVE-2026-19770

The CVE-2026-19770 entry affects feedmob fm-mcp-servers 0.0.3, specifically the Download Endpoint’s Download Report function (src/smadex-reporting/src/index.ts). The vulnerability arises from manipulation of the downloadUrl argument, enabling server-side request forgery (SSRF). Exploitation is de...

EUVD-2026-58563

A vulnerability was identified in feedmob fm-mcp-servers 0.0.3. Affected by this vulnerability is the function downloadReport of the file src/smadex-reporting/src/index.ts of the component Download Endpoint. The manipulation of the argument downloadUrl leads to server-side request forgery. The...

CVE-2026-19770

A vulnerability was identified in feedmob fm-mcp-servers 0.0.3. Affected by this vulnerability is the function downloadReport of the file src/smadex-reporting/src/index.ts of the component Download Endpoint. The manipulation of the argument downloadUrl leads to server-side request forgery. The...

CVE-2026-19244

A vulnerability was detected in HKUDS nanobot up to 0.2.1. The affected element is the function connectmcpservers of the file nanobot/agent/tools/mcp.py of the component MCP enabledTools Scope Handler. Performing a manipulation results in improper access controls. The attack is possible to be...

CVE-2026-19244 HKUDS nanobot MCP enabledTools Scope mcp.py connect_mcp_servers access control

A vulnerability was detected in HKUDS nanobot up to 0.2.1. The affected element is the function connectmcpservers of the file nanobot/agent/tools/mcp.py of the component MCP enabledTools Scope Handler. Performing a manipulation results in improper access controls. The attack is possible to be...