🌐 Reference Implementations for Secure & Compliant AI Workflows with 🔐 AIShield Guardian

Learn more about AIShield Guardian here

This repository provides all integrations leveraging AIShield Guardian for secure, policy-compliant AI workflows for Agentic AI systems.

📜 To know more about 🤖 Agentic AI system and 🛡️AI Agentic Security read this primer.

⚙️ Common Components

• Guardian API : The core GuardianAPIClient class enables input/output validation for compliance in workflows. In case of socket based connectivity use GuardianAPIClient. For testing connectivity leverage the utility of GuardianConnectivityTest.
  • 📝 NOTE: For getting your GUARDIAN_API_ENDPOINT and GUARDIAN_API_KEY, please Contact Us.

🤖 Agentic AI Implementations

• CrewAI : GuardedCrew implementation with example of project planning crew
• Phidata : GuardedAgentWorkflow implementation with the financial agent workflow
• Microsoft AutoGen : MonitorRoundRobinGroupChat implementation with travel planning agents
• OpenAI Agents SDK : input_guardrail/output_guardrail implementation with customer support agent for bank
• PydenticAI : To BE ADDED

🖥️↔️💻 MCP Server Implementation

• MCP : MCPServer implementation with example MCP server and client

🦜️🔗 Langchain Implementation

• Ollama : GuardedOllamaChain implementation with example of chatbot chain

📅 Schedule a call to Learn More!

---

⚖️ License

This project is licensed under MIT License.