The Vulnerable MCP Project

A comprehensive database of Model Context Protocol vulnerabilities, security research, and exploits

50 Vulnerabilities
13 Critical
32 Researchers

Vulnerability Timeline

Loading vulnerabilities...

When a single McpServer instance with a StreamableHTTPServerTransport is reused across multiple clients, responses can leak across client boundaries. One client may receive data intended for another client. Affects MCP TypeScript SDK v1.10.0-1.25.3. CVSS 7.1.

Impact 7/10
Exploitability Moderate
Prevalence Common
Cross-Client Data Leak Security Bypass Data Exfiltration
February 4, 2026 CVE Feed

MCPJam inspector (versions <= 1.4.2) listens on 0.0.0.0 by default with no authentication on its critical endpoint. A crafted HTTP request can install an MCP server and execute arbitrary code on the host. No user interaction required. CVSS 9.8.

Impact 10/10
Exploitability Trivial
Prevalence Emerging
Remote Code Execution Code Execution Security Bypass
February 1, 2026 GitLab Advisory

Three chained vulnerabilities in Anthropic's own mcp-server-git: CVE-2025-68145 (path validation bypass), CVE-2025-68143 (unrestricted git_init that can turn .ssh into a git repo), and CVE-2025-68144 (argument injection in git_diff). Combined with the Filesystem MCP server, these achieve full remote code execution via malicious .git/config files.

Impact 10/10
Exploitability Moderate
Prevalence Common
Command Injection Remote Code Execution Path Traversal Code Execution
January 20, 2026 Security Researchers (The Register)

Remote code execution vulnerability in the GitHub Kanban MCP server. Allows arbitrary command execution through the MCP tool interface.

Impact 9/10
Exploitability Easy
Prevalence Emerging
Command Injection Remote Code Execution Code Execution
January 20, 2026 SentinelOne

Unpatched SSRF vulnerability in the Microsoft MarkItDown MCP server can compromise AWS EC2 instances via metadata service exploitation. The server fetches arbitrary URLs without validation, enabling access to cloud infrastructure credentials. Microsoft classified this as low-risk despite demonstrated EC2 metadata access.

Impact 8/10
Exploitability Easy
Prevalence Emerging
Server-Side Request Forgery Security Bypass Credential Theft
January 20, 2026 David Onwukwe (BlueRock)

The Zen MCP server's is_dangerous_path() function uses exact string matching against a blacklist of sensitive paths, allowing trivial bypass via subdirectory traversal (e.g., /etc/shadow/../../../home/user/.ssh/). Enables reading arbitrary files including SSH keys and API credentials. CVSS 6.5-9.8.

Impact 8/10
Exploitability Easy
Prevalence Emerging
Path Traversal Security Bypass Credential Theft
January 15, 2026 SentinelOne

Path traversal vulnerability in the filesystem-mcp module allows escaping configured directory boundaries to read or write arbitrary files on the host system.

Impact 7/10
Exploitability Easy
Prevalence Emerging
Path Traversal Security Bypass Data Exfiltration
January 15, 2026 SentinelOne

The gemini-mcp-tool package passes unsanitized user input to execAsync shell commands. Network-exploitable with no authentication or user interaction required. CVSS 9.8. As of disclosure, no patch is available (zero-day).

Impact 10/10
Exploitability Trivial
Prevalence Emerging
Command Injection Remote Code Execution Code Execution
January 9, 2026 SentinelOne / Zero Day Initiative

The is_ip_private() function in mcp-fetch-server versions <= 1.0.2 fails to properly validate private IP addresses, allowing Server-Side Request Forgery (SSRF) attacks that reach internal network services. CVSS 9.3.

Impact 9/10
Exploitability Easy
Prevalence Common
Server-Side Request Forgery Security Bypass
December 9, 2025 Team off-course (K-Shield.Jr) / Snyk

Three distinct attack classes exploiting MCP's sampling capability: (1) covert tool invocation to perform hidden file and system operations, (2) conversation hijacking to inject persistent instructions across sessions, and (3) resource theft to drain AI compute quotas for unauthorized workloads.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Prompt Injection Hidden Instructions Resource Theft / Compute Abuse
December 5, 2025 Palo Alto Unit 42

Academic analysis of 67,057 MCP servers across 6 public registries found that a substantial number can be hijacked due to lack of vetted submission processes. Untrusted servers can exfiltrate data from co-connected trusted servers through the shared agent context.

Impact 7/10
Exploitability Moderate
Prevalence Widespread
Supply Chain Data Exfiltration Registry Hijacking
October 18, 2025 Academic Research (arXiv)

The Kluster Verify MCP server allows attackers to drain verification credits through unauthorized access to its MCP tools, causing financial impact and denial of service for legitimate verification operations.

Impact 5/10
Exploitability Easy
Prevalence Rare
Resource Theft / Compute Abuse Security Bypass
October 16, 2025 mcpsec.dev

A case-sensitivity bug in Cursor AI IDE allows attackers to bypass file protection mechanisms and modify .cursor/mcp.json on case-insensitive filesystems (Windows/macOS), enabling injection of malicious MCP servers and achieving remote code execution.

Impact 8/10
Exploitability Moderate
Prevalence Common
Remote Code Execution Security Bypass Code Execution
October 10, 2025 Lakera AI

DNS rebinding vulnerability in the Vet MCP server's SSE transport allows external websites to interact with the locally running MCP server, bypassing browser Same-Origin Policy protections.

Impact 7/10
Exploitability Moderate
Prevalence Emerging
DNS Rebinding Security Bypass
October 6, 2025 mcpsec.dev

Prompt injection attacks against the Amp AI Agent can exfiltrate API keys from the agent's environment. Malicious content processed by the agent extracts credentials and sends them to attacker-controlled endpoints.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Prompt Injection Credential Theft Data Exfiltration
October 3, 2025 mcpsec.dev

Supply chain attack targeting Kilo Code AI Agent users via prompt injection embedded in upstream dependencies. Malicious instructions in package metadata or code comments manipulate the agent's behavior when processing the compromised dependency.

Impact 7/10
Exploitability Moderate
Prevalence Emerging
Supply Chain Prompt Injection
October 2, 2025 mcpsec.dev

Command injection in the Framelink Figma MCP server's fetch-with-retry.ts module. When the standard fetch fails, the server falls back to executing curl via child_process.exec without sanitizing the URL, enabling arbitrary command execution. Over 600,000 downloads and 10,000+ GitHub stars. CVSS 8.0.

Impact 8/10
Exploitability Moderate
Prevalence Common
Command Injection Remote Code Execution Code Execution
October 1, 2025 Imperva Threat Research

The microsoft_docs_fetch tool in the Microsoft Learn MCP Server lacks URL validation, allowing requests to any host instead of restricting to microsoft.com domains. This enables SSRF attacks through the MCP server.

Impact 7/10
Exploitability Easy
Prevalence Emerging
Server-Side Request Forgery Security Bypass
September 19, 2025 Tenable Research

DNS rebinding attack against Coder's Agent API exposes user chat history to attackers. A malicious website can pivot to the locally running Coder agent and read all conversation data.

Impact 6/10
Exploitability Moderate
Prevalence Emerging
DNS Rebinding Data Exfiltration Conversation History
September 19, 2025 mcpsec.dev

DNS rebinding vulnerability in the Neo4j MCP Cypher server (versions 0.2.2-0.3.1) bypasses the browser's Same-Origin Policy, enabling unauthorized tool invocations and complete database takeover of locally running Neo4j instances. CVSS 7.4.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
DNS Rebinding Security Bypass Data Exfiltration
September 11, 2025 GitHub Security Advisory

An attacker shares a malicious Google Doc containing embedded prompt injection. When Cursor fetches the document via a Google Docs MCP server, it auto-executes the injected instructions with no user interaction required. This achieves zero-click remote code execution, credential theft, and persistent access.

Impact 10/10
Exploitability Moderate
Prevalence Emerging
Prompt Injection Remote Code Execution Data Exfiltration Credential Theft +1
September 5, 2025 Lakera AI

The thirdweb MCP server enables unauthorized cryptocurrency transactions through its MCP integration. Attackers can exploit the server to initiate transactions from connected wallets without proper authorization.

Impact 9/10
Exploitability Moderate
Prevalence Emerging
Security Bypass Credential Theft
September 3, 2025 mcpsec.dev

The Grafana MCP server binds to 0.0.0.0:8000 by default, exposing an unauthenticated SSE interface on all network interfaces. Attackers on the network can create, update, and delete Grafana dashboards remotely without any credentials.

Impact 8/10
Exploitability Trivial
Prevalence Emerging
Security Bypass Data Exfiltration
September 2, 2025 mcpsec.dev

Malicious Jira tickets contain obfuscated prompt injection that tricks Cursor into leaking JWT tokens and credentials through the Jira MCP server. A zero-click attack where simply having the agent process a Jira ticket triggers credential exfiltration.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Prompt Injection Credential Theft Data Exfiltration Toxic Agent Flow
August 20, 2025 Snyk Labs

Cursor IDE (versions <= 1.2.4) trusts previously approved MCP configurations indefinitely without re-approval on modification. An attacker adds a benign MCP config to a shared repository, waits for user approval, then replaces it with a malicious payload. This enables persistent remote code execution. CVSS 7.2-8.8.

Impact 8/10
Exploitability Moderate
Prevalence Common
Configuration Bypass Remote Code Execution Bait and Switch
August 1, 2025 Check Point Research

A malicious MCP server can inject JavaScript into Cursor's built-in browser. Cursor lacks integrity checks on runtime components loaded through MCP interactions, enabling arbitrary code execution within the IDE's browser context.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Code Execution Security Bypass Browser Injection
July 22, 2025 0xrelogic (Security Researcher)

All MCP server output vectors—return values, error messages, metadata, resource content, and logging—can carry hidden prompt injection payloads. This demonstrates that no output channel from an MCP server is safe from injection.

Impact 8/10
Exploitability Easy
Prevalence Widespread
Prompt Injection Hidden Instructions Tool Manipulation
July 18, 2025 CyberArk Threat Research

Both the official MCP TypeScript SDK (< 1.24.0) and Python SDK (< 1.23.0) lack DNS rebinding protection for localhost-bound SSE and StreamableHTTP servers, allowing malicious websites to pivot to local MCP servers through the browser.

Impact 8/10
Exploitability Moderate
Prevalence Widespread
DNS Rebinding Security Bypass
July 10, 2025 Security Researchers (GitLab Advisory)

The mcp-remote proxy (versions 0.0.5 through 0.1.15) allows full remote code execution on the client OS when connecting to an untrusted MCP server. A malicious server can craft an authorization_endpoint response that triggers OS command injection. CVSS 9.6. First documented instance of full RCE on the client OS from a remote MCP server.

Impact 10/10
Exploitability Easy
Prevalence Common
Command Injection Remote Code Execution Code Execution
July 9, 2025 JFrog Security Research

The mcp-server-kubernetes package (versions <= 2.4.9) contains unsanitized input in execSync calls within the kubectl_scale, kubectl_patch, and explain_resource tools. Shell metacharacters in tool parameters allow arbitrary command execution on the host. CVSS 7.5.

Impact 8/10
Exploitability Easy
Prevalence Common
Command Injection Code Execution Remote Code Execution
July 8, 2025 GitHub Security Advisory

The node-code-sandbox-mcp package (versions <= 1.2.0) has command injection vulnerabilities that bypass Docker sandbox protections via unsanitized execSync calls. Exploitable through indirect prompt injection when generating code. CVSS 7.5.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Command Injection Sandbox Escape Code Execution
July 8, 2025 GitLab Advisory

Attackers inject a message into a Heroku-hosted web service to trick the MCP-connected agent into transferring ownership of the Heroku application to the attacker. Demonstrates prompt injection through application-level data affecting infrastructure management.

Impact 7/10
Exploitability Moderate
Prevalence Emerging
Prompt Injection Ownership Transfer Tool Manipulation
June 30, 2025 Tramlines.io

The create-mcp-server-stdio package uses unsafe exec() that directly concatenates user input into shell commands, enabling arbitrary command execution on the host system.

Impact 10/10
Exploitability Easy
Prevalence Emerging
Command Injection Remote Code Execution Code Execution
June 20, 2025 GitHub Security Advisory

Critical RCE vulnerability (CVSS 9.4) in the MCP Inspector developer tool (versions below 0.14.1). Lack of authentication between the Inspector client and proxy enables remote code execution, chainable with DNS rebinding for browser-based exploitation.

Impact 10/10
Exploitability Moderate
Prevalence Emerging
Remote Code Execution DNS Rebinding Code Execution
June 13, 2025 Oligo Security

A vulnerability in the GitHub MCP integration where attackers plant prompt injection payloads in public repository issues. When an agent (e.g., Claude Desktop) reviews these issues, it can be manipulated to leak data from private repositories accessible via the same GitHub token.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Toxic Agent Flow Data Exfiltration Prompt Injection
May 26, 2025 Invariant Labs

Malicious MCP servers define tools with parameters like 'system_prompt' or 'conversation_history' that instruct the LLM to populate them with sensitive context. This extracts chain-of-thought, previous tool results, system prompts, and conversation data through seemingly normal tool invocations.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Parameter Abuse Data Exfiltration Hidden Instructions
May 15, 2025 HiddenLayer

Widespread practice across the MCP ecosystem of storing long-term API keys in plaintext on the local filesystem, often with world-readable permissions. Observed in official and third-party MCP servers connecting to services like GitLab, Postgres, and Google Maps.

Impact 8/10
Exploitability Easy
Prevalence Widespread
Credential Theft Plaintext Storage Supply Chain
April 30, 2025 Keith Hoodlet (Trail of Bits)

Attackers use ANSI terminal escape codes to hide malicious instructions in MCP tool descriptions and outputs, rendering them invisible on screen while the LLM processes them normally. Demonstrated against Claude Code with no filtering or sanitization.

Impact 8/10
Exploitability Easy
Prevalence Emerging
Terminal Injection Hidden Instructions Supply Chain
April 29, 2025 Keith Hoodlet (Trail of Bits)

Malicious MCP servers inject trigger phrases into tool descriptions that instruct the LLM to forward entire conversation histories when users type common phrases like 'thank you.' The attack enables passive, long-term data collection.

Impact 8/10
Exploitability Moderate
Prevalence Common
Data Exfiltration Conversation History Trigger Phrase
April 23, 2025 Keith Hoodlet (Trail of Bits)

Malicious MCP servers inject prompts through tool descriptions that manipulate AI behavior immediately upon connection—before any user approval or tool invocation occurs. This bypasses all security checkpoints designed to protect users.

Impact 9/10
Exploitability Easy
Prevalence Widespread
Prompt Injection Hidden Instructions Security Bypass
April 21, 2025 Trail of Bits

A sophisticated two-stage attack combining tool poisoning with remote code execution. The first stage creates a persistence marker; the second stage modifies the tool's docstring to include social engineering content that tricks AI assistants into executing base64-encoded commands for SSH key theft.

Impact 10/10
Exploitability Moderate
Prevalence Emerging
Remote Code Execution SSH Key Theft Base64 Obfuscation Data Exfiltration
April 17, 2025 Repello AI

MCP's stdio transport enables frictionless local server execution, creating a low-barrier path for users to download and run potentially malicious third-party code on their machines with full local privileges.

Impact 6/10
Exploitability Easy
Prevalence Common
Code Execution Supply Chain Social Engineering
April 13, 2025 Shrivu Shankar

MCP tools can silently mutate their own definitions after initial user approval. A tool that appears safe during installation can later change its behavior to perform malicious actions without notifying the user.

Impact 7/10
Exploitability Easy
Prevalence Common
Tool Manipulation Bait and Switch Security Bypass
April 9, 2025 Elena Cross (via Simon Willison)

When multiple MCP servers are connected to the same agent, a malicious server can override or intercept calls intended for a trusted server by registering tools with the same or similar names.

Impact 7/10
Exploitability Moderate
Prevalence Common
Tool Hijacking Confused Deputy Tool Overriding
April 9, 2025 Elena Cross (via Simon Willison)

An attack against the whatsapp-mcp server where a malicious co-installed server steals WhatsApp message history by swapping its tool definition after approval and using UI deception techniques to hide the exfiltrated data.

Impact 8/10
Exploitability Moderate
Prevalence Emerging
Data Exfiltration Message Hijacking Deceptive UI
April 7, 2025 Invariant Labs

A peer-reviewed study systematically demonstrating that LLMs with MCP access enable major security exploits including code execution, remote access control, and credential theft. Introduces the MCPSafetyScanner open-source audit tool for assessing MCP server security.

Impact 5/10
Exploitability Theoretical
Prevalence Widespread
Supply Chain Code Execution Credential Theft
April 2, 2025 Academic Research (arXiv)

A specialized form of prompt injection where malicious instructions are embedded in tool descriptions—visible to the LLM but hidden from users. Attackers create tools with concealed directives that cause the LLM to perform unauthorized actions such as exfiltrating private data.

Impact 9/10
Exploitability Easy
Prevalence Widespread
Data Exfiltration Prompt Injection Hidden Instructions
April 1, 2025 Invariant Labs

Malicious MCP servers trigger repeated consent requests to fatigue users into granting excessive permissions, exploiting human psychology to bypass approval-based security controls.

Impact 6/10
Exploitability Easy
Prevalence Common
Social Engineering Consent Bypass Security Bypass
April 1, 2025 Palo Alto Networks

MCP servers with identical tool names create naming collisions where malicious servers can override legitimate ones. Without namespacing, the behavior depends on client implementation and connection order.

Impact 6/10
Exploitability Moderate
Prevalence Common
Naming Collision Tool Overriding Confused Deputy
April 1, 2025 Various (Embracethered blog)

The MCP protocol specification places session identifiers in URLs (e.g., GET /messages/?sessionId=UUID), violating security best practices and exposing session tokens in logs, browser history, and referrer headers.

Impact 6/10
Exploitability Trivial
Prevalence Widespread
Session Hijacking Protocol Design
March 29, 2025 Equixly