AI App with both MCP and A2A

 

Service Now AI Agent

Business App ( CMDB )  - will also take care heigene 

ensure A2A protocol working

 Workflow

KB articles as Agents

Tool 

MCP Vs A2A

 

A2A and the Model Context Protocol (MCP) are complementary standards for building robust agentic applications:

• MCP (Model Context Protocol): Connects agents to tools, APIs, and resources with structured inputs/outputs. Think of it as the way agents access their capabilities.
• A2A (Agent2Agent Protocol): Facilitates dynamic, multimodal communication between different agents as peers. It's how agents collaborate, delegate, and manage shared tasks.

langGrpah Open Agent Framework

 

Highlights

• No different from Azurre AI Foundry or AWS bedrock or GCP Agent garden
• Supports only parallel execution of Tasks
• No suppor for Context Memory
• Suports native connectors
• Available in Private cloud

Information Systems Security Agent ( Multi Agent)

Agentic AI-powered security automation framework that performs:

1. Threat modeling (using STRIDE, OWASP Top 10),

2. Risk assessment (using DREAD/FAIR models),

3. Architecture parsing (from diagrams or IaC),

4. Compliance mapping (to NIST, ISO, SOC2),

5. Live Azure infrastructure auditing (e.g., VNet, NSG, Key Vaults, Route Tables),

6. Automated remediation planning (with suggested fixes),

7. Audit-ready reporting (in PDF or dashboard format),

8. ServiceNow integration (for CMDB, tickets, approvals).

Build this using a multi-agent system (LangGraph or Azure AI Foundry Agents) coordinated by a Supervisor Agent, with contextual memory (e.g., Azure AI Search or Weaviate). Include:

• A system architecture diagram,

• A step-by-step description of each agent's function and AI implementation,

• An ROI analysis per agent (quantifying time/money saved),

• A PowerPoint presentation summarizing all components,

• A PNG diagram of the architecture, and

• Exportable formats (PowerPoint, draw.io, or PlantUML if needed).