Glossary

Dictionary of terms for the Enterprise Multi-Agent Team framework.

A B C D E G L M O P R S T W

A

A2A Protocol protocol

Agent-to-Agent Protocol. A standardized protocol for communication between AI agents, enabling task delegation, result sharing, and collaborative workflows across different agent frameworks.

agent.send_task(task_id="task-001", message="Analyze data")

See also: Agent, Workflow

Agent noun

An autonomous AI entity programmed to perform specific tasks using LLM capabilities. Agents can use tools, execute functions, and communicate via A2A Protocol.

agent = OpenAIAgent() result = agent.run("Write documentation")

See also: SDK, LLM

API abbreviation

Application Programming Interface. A set of protocols and tools for building software applications. The Enterprise Team provides a REST API for agent management.

POST /agents/{agentId}/run

B

Base Agent noun

The foundational abstract class that all agent implementations inherit from. Provides common interfaces for running tasks, listing capabilities, and registering skills.

class MyAgent(BaseAgent): ...

C

Capability noun

A skill or function that an agent can perform. Defined in the agent card with name, description, and input/output schemas.

Capability(name="code_execution", description="Execute code")

See also: Skill, Agent Card

Coordinator noun

A special agent type that orchestrates multiple agents to work collaboratively on complex tasks.

team = TeamCoordinator() team.run_parallel(agents, task)

See also: Workflow

D

DID abbreviation

Decentralized Identifier. A format for uniquely identifying resources in the registry (e.g., did:agent:openai).

did:agent:openai

G

GitHub Actions framework

A CI/CD platform for automating build, test, and deployment workflows.

uses: actions/checkout@v4

L

LangGraph framework

A library for building stateful, multi-agent applications powered by LLMs. Used for workflow orchestration in this project.

workflow = create_workgraph() workflow.invoke(state)

See also: Workflow

LLM abbreviation

Large Language Model. An AI model trained on vast amounts of text data, capable of generating human-like text.

client = create_llm_client("openai", model="gpt-4o")

See also: Provider

M

Message noun

In A2A Protocol, a structured communication containing role, parts, and metadata between agents.

Message(role="user", parts=[TextPart(text="Hello")])

O

OpenAPI protocol

A specification for building and documenting REST APIs. The Enterprise Team API follows OpenAPI 3.0.

openapi: 3.0.3
Orchestrator noun

A component that manages the execution flow of multiple agents using LangGraph.

orchestrator.run(team, task)

P

Provider noun

An LLM service provider (OpenAI, Anthropic, Google, etc.) that can generate text.

client = create_llm_client("anthropic")

See also: LLM

R

Registry noun

A centralized store for skills, prompts, and tools using DID format for identification.

registry = get_registry() skills = registry.list_skills()

S

SDK abbreviation

Software Development Kit. A library for building applications for a specific platform or service.

from openai import OpenAI

See also: Agent

Skill noun

A capability that can be invoked on an agent. Registered with ID, name, and description.

Skill(id="coding", name="Coding", description="Software dev")

See also: Capability

T

Task noun

In A2A Protocol, an execution of work containing ID, message, and status information.

Task(id="task-001", message=message, status=status)

W

Workflow noun

A directed graph of agent execution steps orchestrated by LangGraph. Supports sequential, parallel, and hierarchical strategies.

workflow.invoke({"task": "Write content"})

See also: LangGraph, Coordinator