Security MCP servers for scanning, threat intelligence and audits: let an AI assistant check code, dependencies and infrastructure for risks. 3,538 servers in this category.
Scans MCP servers and agent tool manifests for poisoning, exfiltration, secrets, and supply-chain risks.
$claude mcp add palanit-core -- uvx palanit-core
A transparent proxy that independently re-verifies MCP tool-call claims and prevents duplicate side effects through idempotent replays.
$claude mcp add verimcp -- uvx verimcp
Performs evidence-first security reviews of authorized repositories by mapping attack surfaces and verifying vulnerabilities through independent verification.
$claude mcp add sechelix -- uvx sechelix
Maps every service that sends email for a given domain using public DNS, certificate transparency logs, and website records.
$claude mcp add sendermap -- uvx sendermap
Injects API keys into command execution environments without exposing secret values to the AI model's context.
$claude mcp add keygrant -- uvx keygrant
Detects behavioral drift in Claude Agent Skills and MEMORY.md files by performing static analysis on declared and inferred capabilities.
$claude mcp add evolveguard -- uvx evolveguard-cli
Verifies code integrity by detecting fabricated success paths, hollow tests, and security vulnerabilities like BOLA/IDOR.
$claude mcp add prufgate -- uvx prufgate
Flags deceptive signals and security risks in AI-agent marketplace listings, including prompt injection and fake engagement.
$https://github.com/Echolonius/boardcheck/releases/download/v0.8.1/boardcheck-mcp-0.8.1.mcpb
Provides AI agents with durable identity and memory by anchoring data on the Emercoin blockchain.
$claude mcp add gateway -- docker run -i --rm ghcr.io/steledger/mcp:latest
Lets an AI client perform URL scanning, malware detection, and security analysis using the ScanMalware.com API.
$claude mcp add --transport http mcp-server https://mcp.scanmalware.com/mcp
Secures AI agents by auditing MCP configurations, redacting secrets, vetting skills, and performing network scans.
$claude mcp add securitystack-mcp -- uvx securityscan-mcp
Investigate domains, IPs, and websites using public records, certificate transparency logs, and infrastructure comparisons.
$claude mcp add osint-mcp -- uvx osint-mcp-server
Enables AI agents to interact with the CrowdStrike Falcon platform for security operations, threat intelligence, and incident response.
$claude mcp add falcon-mcp -- uvx falcon-mcp-extended
Allows an AI client to execute read-only SQL queries on a PostgreSQL database using syntax tree parsing to prevent unauthorized write operations.
$claude mcp add pg-readonly-mcp -- uvx pg-readonly-mcp
Lets an AI assistant access the Internet Cyber Health Index to monitor global cyber threat bands, scores, and critical advisories.
Provides read-only breach intelligence by querying public disclosure feeds to report on organization breach history, scale, and data types.
$claude mcp add data-breach-detector -- uvx data-breach-detector
Analyzes Python and TypeScript agent applications for architecture quality, finding missing execution bounds, resilience controls, and observability.
$claude mcp add uptocode -- uvx uptocode
Enables secure transactional email sending through a two-phase preview and confirmation process with built-in rate limits and recipient allowlists.
$claude mcp add sendgrid-mcp-secure -- uvx sendgrid-mcp-secureA stdio proxy that automatically logs tool calls and results to a hash-chained ledger to ensure an immutable record of MCP server interactions.
$claude mcp add adapter -- uvx arcaeon-adapter
Provides a production-grade gateway for MCP servers with OAuth security, PII scrubbing, and a governed LangGraph agent for enterprise environments.