Security MCP servers for scanning, threat intelligence and audits: let an AI assistant check code, dependencies and infrastructure for risks. 3,537 servers in this category.
Provides a local-first security layer for AI agents to review shell commands, file access, package installs, and MCP tool calls before execution.
$claude mcp add hol-guard -- uvx hol-guard
Runs coding agent commands, file reads, and writes on a remote host over SSH while keeping credentials and the local machine closed.
$claude mcp add farhand -- npx -y farhand
Lets an AI client open binaries in headless IDA Pro to analyze functions, decompile code, and edit names, types, and comments.
Provides long-term memory and a safety gate for AI coding agents using signed logs and rule-based tool call blocking.
$claude mcp add vestige -- npx -y vestige-mcp-server
Lets an AI agent create, inspect, and manage secure tunnels that expose local HTTP, TCP, UDP, and P2P services to the public internet.
$https://github.com/joaoh82/rustunnel/releases/download/v0.8.9/rustunnel-mcp-aarch64-apple-darwin.mcpb
Lets an AI client explore and query MySQL, MariaDB, PostgreSQL, and SQLite databases with PII redaction and read-only defaults.
$claude mcp add dbmcp -- docker run -i --rm ghcr.io/haymon-ai/dbmcp:0.18.1
Runs sandboxed WebAssembly plugins distributed via OCI registries to give AI clients extra tools like time, hashing, QR codes, and web fetch.
Provides a secure, read-only interface for AI agents to query PostgreSQL databases using AST validation and cost-based query guards.
$claude mcp add postgres-mcp-hardened -- npx -y postgres-mcp-hardened
Analyzes application code to generate baseline AWS IAM identity policies for AI coding assistants.
Acts as a local credential broker that injects secrets into AI agent requests via scoped presets to prevent sensitive data from entering context windows.
$mcp-warden
Provides cryptographic signing, verification, and hash-chained audit trails for AI agent tool calls to ensure tamper-evident evidence.
$claude mcp add signet -- npx -y @signet-auth/mcp-tools
Acts as a secure gateway between AI editors and MCP servers, providing sandboxing, secret management, and centralized tool orchestration.
$https://github.com/Soflutionltd/McpHub/releases/download/v5.9.1/mcphub.mcpbActs as a secure central gateway for MCP servers, providing sandboxing, secret management, and optimized tool definitions for AI editors.
$https://github.com/Soflutionltd/McpHub/releases/download/v5.11.0/mcphub.mcpb
Provides AI agents with secure, token-efficient, and asynchronous SSH access to remote servers using a zero-trust whitelist approach.
$agentic_ssh
Redacts sensitive credentials and prunes framework noise from error logs to optimize context for AI coding agents.
$tokenectomy
Redacts sensitive credentials and prunes framework noise from error logs to reduce token usage and prevent AI hallucinations.
$claude mcp add tokenectomy -- npx -y tokenectomy-razor
Allows an AI client to manage Palo Alto Networks PAN-OS firewalls through scoped, audited, and secure network-security automation.
$claude mcp add rustpanosmcp -- docker run -i --rm ghcr.io/mechubsec/rustpanosmcp:0.17.1
A Rust-based security sidecar that protects MCP servers from prompt injection, tool poisoning, and Unicode evasion attacks.
$claude mcp add mcp-armor -- docker run -i --rm ghcr.io/studiomeyer-io/mcp-armor:0.8.2
Performs evidence-aware CVE research by aggregating data from NVD, EPSS, CISA KEV, and OSV to scan SBOMs, lockfiles, and container images.
$oxiploitability-mcp
Lets an AI client index local codebases and run semantic search, call graph, taint and supply chain analysis across 32 languages.