Security MCP servers for scanning, threat intelligence and audits: let an AI assistant check code, dependencies and infrastructure for risks. 3,541 servers in this category.
Performs static security triage on MCP server configurations to identify secrets, dangerous tool exposures, and prompt injection risks.
Maps OWASP MCP Top 10 risks to NIST SP 800-53 and CSF 2.0 controls and enriches SARIF security scan outputs with these compliance mappings.
Provides a permission-tier gate that blocks outbound agent tool executions until a human provides explicit approval.
Lets an AI assistant scan and audit home local area networks by using an Android phone as a secure bridge and patrol device.
Governs AI agents by routing model calls through an LLM gateway, enforcing read-only tool access, masking sensitive data, and scoping user permissions.
Enforces security boundaries for AI agents by validating tool calls against allowlists, fingerprints, scopes, argument rules, and egress controls.
Scans infrastructure for AI agent exposure, including unauthenticated MCP servers, hidden instructions, and open AI services like Ollama.
Analyzes MCP servers by executing them in an isolated Docker sandbox to confirm and reproduce implementation vulnerabilities like SSRF and command injection.
Tests MCP servers for vulnerabilities by sending malformed JSON-RPC payloads to identify crashes, hangs, and information leaks.
Performs offline, read-only security audits of MCP configuration files to detect credential leaks and tool-description poisoning.
Probes OAuth discovery issuer validation to prevent credential leakage to attacker-controlled token endpoints.
Provides a secure, governed entry point for AI agents to access enterprise tools and other agents through pinned tool definitions and signed audit chains.
Securely expose local MCP servers to Claude.ai using Cloudflare Tunnel and Access with origin-side JWT re-verification.
Provides a link-safety gate for AI agents by using remote browsers to inspect suspicious links and provide evidence before they are opened.
Deploys hardened, multi-tenant MCP servers on Cloud Run with Google identity verification, Firestore rate limiting, and BigQuery auditing.
Benchmarks static analyzers on real, publicly disclosed vulnerabilities in open-source MCP servers using Docker and SARIF scoring.
Implements Prompt State Protocol and Covenant Declaration Language to secure LLM workflows through signed prompts and data classification policies.
Scans MCP tool declarations for security risks like command injection, secret exposure, and improper tenant boundaries.
Provides a secure local broker for agents to request secrets that are pasted into the user's OS keystore without the agent ever seeing the actual value.
Scans networks to discover unregistered MCP servers and classifies their security posture and available tools.
