Security proxy for AI agents. Scans every message for prompt injection, PII, and secrets. Defense-in-depth: Go proxy + iptables firewall + eBPF kernel monitor. YAML policy engine, audit logging, 5 AI agents with RAG knowledge bases.

go reverse-proxy ebpf ai-agents vulnerability-scanning ai-security audit-logging defense-in-depth pii-detection prompt-injection llm-security security-proxy openclaw

• Updated Mar 26, 2026
• Go

A high-performance Zero-Trust security proxy designed for data privacy and compliance. Implements dynamic data masking (DDM), AES-256 encryption, and automated PII detection to secure sensitive data flows in enterprise environments. Engineered for HIPAA/GDPR alignment.

python cryptography encryption cybersecurity data-privacy zero-trust api-security compliance-as-code fastapi backend-engineering hipaa-compliance security-proxy

• Updated Jun 30, 2026
• Python

🛡️ A firewall for your agent's brain. Open-source security proxy that protects AI agents from prompt injection attacks. 4-stage pipeline, community threat feed, no bypass mode.

open-source raspberry-pi firewall oled ai-agents ai-security prompt-injection llm-security security-proxy openclaw

• Updated Apr 26, 2026
• JavaScript

Open-source security gateway for MCP agents and tools. Inspect tool calls before execution, enforce policy, block risky operations, and emit audit-ready evidence.

python open-source mcp firewall sse cybersecurity railway ai-security agentic-ai model-context-protocol anthropic-mcp security-proxy

• Updated Nov 7, 2026
• Python

Reliability & security proxy for the Model Context Protocol (MCP). Self-healing connections, runtime tool-poisoning/shadowing defense, and NIST AI RMF + OWASP LLM Top-10 audit trails for any MCP server. Works with Claude, Cursor, Cline, Windsurf.

typescript ai proxy mcp owasp cursor ai-agents claude ai-security llm prompt-injection anthropic llm-security agentic model-context-protocol mcp-server mcp-proxy nist-ai-rmf security-proxy tool-poisoning

• Updated Nov 9, 2026
• TypeScript

Open source security for LLM inference

open-source openai ai-security prompt-injection anthropic llm-security security-proxy

• Updated Mar 11, 2026
• JavaScript

Security proxy for AI agents (OpenClaw, LangChain, CrewAI). Protects API keys, enforces spend limits, blocks dangerous endpoints. Local-first Rust binary — your credentials never leave your machine.

rust rate-limiting ai-safety local-first credential-vault eip-712 zk-proofs llm-security defi-security crypto-wallet-security ai-agent-security security-proxy openclaw api-key-proxy spend-limits

• Updated Mar 16, 2026
• Rust

Zero-dependency supply-chain security proxy for npm, PyPI, and Cargo. Scores packages against attack patterns at install time.

rust npm security pypi supply-chain ci-cd cargo sca devsecops software-composition-analysis zero-dependencies supply-chain-security package-security security-proxy

• Updated Nov 7, 2026
• Rust

Provenance-aware security proxy for AI agents — block prompt-injection actions at the MCP tool boundary, with no agent code changes.

python security azure mcp cybersecurity ai-agents information-flow-control ai-security azure-openai prompt-injection llm-security agentic-ai model-context-protocol security-proxy

• Updated Nov 7, 2026
• Python

Production-grade, sub-5ms security & rate-limiting reverse proxy for Model Context Protocol (MCP) servers. Features local ONNX prompt injection filtering, Redis Lua rate-limiting, stdio-to-HTTP bridging, and Grafana observability.

docker typescript mcp rate-limiting prometheus redis-lua graphana onnx prompt-injection model-context-protocol security-proxy

• Updated Nov 11, 2026
• TypeScript

A real-time PostgreSQL proxy that anonymizes sensitive data on the fly for secure development

nodejs typescript nextjs postgresql tcp-proxy developer-tools data-privacy database-proxy data-anonymization real-time-monitoring data-masking database-security kvkk gdpr-compliance database-security-lab pii-protection security-proxy

• Updated Jan 23, 2026
• TypeScript

An open-source, wire-transparent security proxy for PostgreSQL that intercepts and blocks destructive SQL queries from AI agents before they hit your database.

go golang open-source postgres devops postgresql rate-limiting python-sdk ai-agents ai-security database-security guardrails llm security-proxy

• Updated Oct 26, 2026
• Go

Transparent policy proxy + structuring detector for Razorpay's MCP server

mcp audit-log razorpay policy-engine agentic-ai financial-security security-proxy

• Updated Oct 22, 2026
• TypeScript

Lightweight, zero‑dependency Node.js proxy that transparently forwards HTTP traffic between local AI agents and healthcare APIs, keeping PHI handling and HIPAA compliance in the downstream services.

nodejs javascript http-proxy healthcare-api hipaa-compliant security-proxy

• Updated Apr 2, 2026

Rule-based LLM security proxy for PII leakage prevention and prompt injection mitigation in public-sector civil service workflows.

python privacy public-sector data-loss-prevention capstone-project fastapi pii-detection prompt-injection llm-security security-proxy

• Updated Nov 2, 2026
• Python

Rust security proxy for MCP servers with taint tracking, policy enforcement, and secret redaction.

rust mcp ai-security tla-plus policy-as-code taint-tracking prompt-injection agentic-ai model-context-protocol security-proxy

• Updated Nov 11, 2026
• Rust

A security-first RAG framework featuring bi-directional PII masking and local embeddings. Designed for secure technical enablement and compliance-aware AI integration in high-discretion enterprise environments.

python data-privacy dlp solutions-engineering faiss rag pydantic vector-database langchain retrieval-augmented-generation enterprise-ai pii-masking security-proxy technical-enablement

• Updated Feb 19, 2026
• Python

A firewall for AI agents — it finds and redacts hidden instructions in the untrusted content an agent reads (emails, web pages, tool results), so they never reach the model's prompt.

security-gateway prompt-injection-llm-security mcp-proxy prompt-injection-defense prompt-injection-detection security-proxy prompt-injections prompt-injection-protection

• Updated Jul 9, 2026
• Go

Dual-layer MCP security proxy — prompt injection filtering, capability attestation, sandboxed code execution, and multi-turn attack detection for multi-server MCP deployments.

python docker security mcp json-rpc sandboxing ai-safety fastapi llm llm-security agentic-ai model-context-protocol security-proxy prompt-inkection

• Updated Oct 7, 2026
• Python