Security & Testing Servidores MCP

Box Pydantic AI

Demonstrates the integration of Pydantic AI agents with the Box MCP server for secure content access.

Agentic

Exposes runtime MCP Server vulnerabilities related to GitHub and Linear integrations.

Playwright

Provides robust browser automation capabilities, enabling large language models to interact with web pages, generate code, and extract data.

FPE Demo

Demonstrates secure format-preserving encryption (FF3 FPE) and multi-mode authentication for sensitive data within LLM interactions.

Term

Provides secure, auditable remote access to a system shell for executing commands and streaming output.

OAuth

Implements a complete OAuth 2.1 server for FastMCP, facilitating robust testing of client authentication flows and token management.

Skill Loader

Manages Claude skills by discovering, fetching, validating, and converting them from skills.sh and GitHub repositories.

YAR

Provides AI assistants with fast, self-hosted access to GitHub repositories, eliminating rate limits and ensuring up-to-date source code availability.

OSSEC

Exposes OSSEC HIDS security monitoring and management capabilities as tools, resources, and prompts for AI assistants.

Errs

Normalizes verbose compiler and linter output into structured JSON for compact, programmatic error analysis.

Truthseeker

Provides ground truth validation for AI agents, preventing hallucinated claims and ensuring data quality.

Defense

Provides AI assistants access to 94 defensive security tools on Linux, enabling system hardening, compliance, and incident response via natural language.

Foglift

Analyzes websites for traditional search engine optimization (SEO), Generative Engine Optimization (GEO), performance, security, and accessibility.

CMD-Line

Enables AI assistants to securely execute terminal commands with controlled directory access and granular permissions.

QEMU Virtual Machine

Enables AI agents to visually interact with, control, and manage QEMU/KVM virtual machines through VNC, QMP, and Guest Agent for comprehensive automation.

SecureCode

Enables AI agents to securely access sensitive secrets without exposing their raw values directly in chat.

Easycrypt

Integrates EasyCrypt proof assistant functionalities into the Claude Code AI assistant, enabling comprehensive interaction with cryptographic proofs.

RTK

Intercepts and compresses shell command output by 60–90% for Claude Desktop, optimizing token usage and preventing prompt injection.

Open Source Supply Chain Risk

Provides AI agents with direct access to critical intelligence for open-source supply chain risk analysis, including CVEs, CISA KEV exploits, bus-factor scores, and typosquat detection.

RegisterOracle

Manages DORA Art. 28 Register of Information for ICT providers and facilitates ITS export.