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.
Masks personally identifiable information in outbound MCP tool-call arguments and restores original values in the results.
$claude mcp add mcp -- uvx maskflow-mcp
Scans MCP servers and AI agents for prompt injection, tool poisoning, supply chain risks, and identity vulnerabilities.
$claude mcp add aishield -- npx -y aishield-mcp-server
Scans machine learning model files like PyTorch, pickle, and GGUF for malicious code that executes upon loading.
$claude mcp add hayward -- uvx hayward
Validates governance content in XML data instances against Semantic Data Charter models using OASIS XACML decision semantics.
$claude mcp add sdcgovernance -- uvx sdcgovernance
Acts as an authorization proxy for MCP servers using dual-path verification through deterministic policies and AI judges to approve tool calls.
$claude mcp add two-key -- uvx two-keyLets AI agents perform identity resolution, platform registration checks, and breach detection using phone numbers and email addresses.
$claude mcp add --transport http outris-identity-mcp https://mcp-server.outris.com/http
Provides AI agents with safe, read-only SQL access to multiple database engines including PostgreSQL, MySQL, Oracle, and SQL Server.
$claude mcp add universal-db-mcp -- docker run -i --rm ghcr.io/alghanim/universal-db-mcp:0.2.0
Provides a security suite for AI agents to detect transaction drains, permit-phishing, and enforce action policies before signing blockchain transactions.
$claude mcp add chain-signer -- uvx chain-signer
Provides deterministic policy evaluation and tamper-evident audit chains for AI agent outputs to ensure safety and verifiability.
$claude mcp add --transport http dcl-webhook https://mcp.fronesislabs.com/mcp
Lets AI shopping agents detect scams, dropshipping, and dark patterns in store or product listings using Solana-based payments.
$claude mcp add mcp -- uvx unfret-mcp
Masks personal data and secrets in AI agent tool calls by routing requests through a secure node that enforces privacy policies.
$claude mcp add vrymdor-node-python -- uvx vrymdor-node
Performs plain-English security scans on AI-built applications and generates fix requests for coding agents.
$claude mcp add appledger -- uvx appledger
Assess the post-quantum cryptography readiness of websites and access global PQC adoption statistics.
$claude mcp add client -- uvx witnessory
Enables secure, account-free authenticated messaging between agents using signed sessions and encrypted channels over a public hub.
$claude mcp add agentcomms -- uvx agentcomms-secure
Provides AI agents with governed, read-only access to DuckDB, Postgres, or BigQuery using SQL guards, row policies, and PII pseudonymization.
$claude mcp add warehouse-mcp -- uvx warehouse-mcp
Acts as an MCP proxy that requires passkey approval for risky tool calls and provides signed, offline-verifiable receipts for agent actions.
$claude mcp add witness -- uvx deus-witness
Provides read-only visibility into Google Workspace security events, including suspicious logins, external file sharing, and account status.
$claude mcp add gwsadm-mcp -- uvx gwsadm-mcp
Allows an AI agent to check the danger grade and transparency of public MCP servers before establishing a connection.
$claude mcp add mcp-trust -- uvx mcp-trust
Provides coding agents with read-only SQL schema knowledge and PII-safe query capabilities for SQL Server, MySQL, and MariaDB.
$claude mcp add sql-safe-mcp -- uvx sql-safe-mcpProvides coding agents with read-only SQL schema knowledge and PII-safe query capabilities for SQL Server, MySQL, and MariaDB.
$claude mcp add sql-mini-mcp -- uvx sql-mini-mcp