agenticx-a2a-connector

AgenticX A2A Connector

Guide for building distributed, inter-communicating agent systems using the A2A protocol.

What is A2A?

A2A (Agent-to-Agent) is a protocol that enables agents to discover each other's capabilities and invoke them as if they were local tools. This allows building distributed agent systems where specialized agents collaborate across network boundaries.

Core Components

Component	Purpose
AgentCard	Advertises an agent's identity and skills
Skill	Describes a capability an agent offers
A2ASkillTool	Wraps a remote skill as a local tool
A2ASkillToolFactory	Batch-creates tools from an AgentCard
A2AClient	HTTP client for calling remote agents

Agent Cards

An AgentCard declares what an agent can do:

from agenticx.protocols import AgentCard, Skill

card = AgentCard(
    name="Research Agent",
    description="Specializes in web research and report generation",
    url="http://research-agent:8000",
    skills=[
        Skill(
            name="web_research",
            description="Search the web and compile findings",
            parameters_schema={
                "type": "object",
                "properties": {
                    "query": {"type": "string"},
                    "depth": {"type": "integer", "default": 3}
                },
                "required": ["query"]
            }
        ),
        Skill(
            name="generate_report",
            description="Generate a structured report from research data",
            parameters_schema={
                "type": "object",
                "properties": {
                    "topic": {"type": "string"},
                    "format": {"type": "string", "enum": ["markdown", "html"]}
                },
                "required": ["topic"]
            }
        )
    ]
)

Using Remote Agent Skills as Tools

Single Skill

from agenticx.protocols import A2ASkillTool

tool = A2ASkillTool(
    agent_url="http://research-agent:8000",
    skill_name="web_research"
)

# Use like any local tool
result = tool.run(query="latest AI trends", depth=5)