Terms
Every term in this hub.
This hub collects the orchestration terms in the Prefactor glossary. The patterns group covers how an individual agent thinks in steps: ReAct, plan-and-execute, planning loops, and the agentic loop itself. These are the concepts you meet first, because every multi-agent system is built out of single agents that use them.
The multi-agent group is where the coordination problems live: orchestrator and supervisor agents, sub-agents, handoffs, and swarms. The routing group covers the decision of which agent or model handles a request at all. And the memory group covers what persists between steps and sessions, because an agent that forgets everything between turns pushes the coordination burden onto everything around it.
A fair warning about this vocabulary: it is the least standardised in the glossary. What one framework calls a supervisor another calls an orchestrator, and the definitions here note those collisions where they matter rather than pretending one vendor's usage won.
Terms are grouped below from the inside out: reasoning patterns first, then multi-agent coordination, routing, memory and state, and finally the infrastructure pieces such as queues and middleware. Each entry links to its full definition.
Reasoning patterns & loops
Agentic Design PatternsAgentic design patterns are the recurring structures teams reuse when building agent systems, much as design patterns describe recurring solutions in code.Agentic LoopAn agentic loop is the core execution cycle of an autonomous agent: observe the environment, reason about the next step, take an action, and evaluate the result.Plan-and-Execute (Agent Pattern)Plan-and-execute is an agent reasoning pattern in which the agent first generates a step-by-step plan for completing a task, then executes each step sequentially or in parallel.Planning LoopA planning loop is the iterative cycle in which an agent breaks a goal into subtasks, selects tools or actions, executes them, evaluates the result, and re-plans if needed.ReAct PatternReAct (Reasoning + Acting) is an agent architecture pattern where the model alternates between reasoning about what to do next and taking an action such as calling a tool.Sidecar Pattern (Agent)The sidecar pattern deploys a governance or observability agent as a co-located process alongside the main AI agent, intercepting its calls to inject logging, policy enforcement, or telemetry without modifying the agent's code.
Multi-agent coordination
Agent DelegationAgent delegation is when one agent assigns a subtask to another agent, transferring responsibility for that portion of the work.Agent HandoffAn agent handoff is the transfer of a task, context, and control from one agent to another during a workflow.Agent OrchestrationAgent orchestration is the coordination of multiple AI agents working together on a shared task.AI Agent Swarm IntelligenceSwarm intelligence in AI agents is when many simple agents collectively solve problems through decentralised coordination: similar to how ant colonies or bird flocks exhibit intelligent behavior.Behavioral OrchestrationBehavioral orchestration is the coordination of how multiple agents act across a workflow.Delegated AccessDelegated access is a pattern in which a user grants an AI agent temporary, scoped permission to act on their behalf.Delegation ControlDelegation control is the governance capability that limits how an AI agent can pass its authority to other agents or services.Multi-Agent CollaborationMulti-agent collaboration is when multiple agents cooperate on one outcome by sharing tasks, context, or results.Multi-Agent SystemA multi-agent system is an architecture in which multiple AI agents collaborate or compete to complete tasks that are too complex for a single agent.Orchestrator AgentAn orchestrator agent is an AI agent responsible for coordinating other agents in a multi-agent workflow, decomposing a high-level goal into sub-tasks, assigning them to specialist agents, managing dependencies between tasks, and aggregating results.Sub-AgentA sub-agent is an agent spawned by a parent agent to handle a specific subtask.Supervisor AgentA supervisor agent is a higher-level agent that delegates subtasks to worker agents, monitors their progress, and decides when to intervene, retry, or escalate.Swarm ArchitectureA swarm architecture is a multi-agent design where many lightweight agents operate in parallel on related subtasks with minimal central coordination.Sybil Attack (Multi-Agent)A sybil attack in a multi-agent system creates many fake or compromised agent identities to gain disproportionate influence over collective decisions, reputation systems, or resource allocation.
Routing
Memory & state
Agent MemoryAgent memory is the mechanism that lets an agent retain information across interactions, either within a session (short-term) or across sessions (long-term).Agentic MemoryAgentic memory refers to the mechanisms by which an AI agent retains and retrieves information across steps within a task or across multiple sessions.Episodic Memory (Agent)Episodic memory in an AI agent is the retention of specific past events, interactions, completed tasks, or decision outcomes, that the agent can retrieve and reason about in future sessions.Long-Term Memory (Agent)Long-term memory in an AI agent system is persistent storage that survives across individual tasks and sessions, allowing agents to recall facts, preferences, and past interactions over time.Semantic Memory (Agent)Semantic memory in an AI agent is the storage of general world knowledge, domain facts, and declarative information that the agent can retrieve when needed.Short-Term Memory (Agent)Short-term memory in an AI agent is task-scoped storage that persists across the steps of a single execution but is discarded when the task ends.
Infrastructure
Agent MeshAn agent mesh is a network architecture where multiple agents discover, communicate with, and invoke each other through a shared protocol layer, similar to a service mesh in microservices.Agentic WorkflowAn agentic workflow is a multi-step process where one or more AI agents make decisions, call tools, and pass results between stages to reach an outcome.Message Queue (Agent)A message queue in an agent architecture is a durable buffer that decouples the producers of agent tasks from the agents that process them, enabling asynchronous execution, load levelling, and guaranteed delivery.
More agent orchestration & workflow terms
Agent FabricAn agent fabric is the underlying infrastructure and middleware that enables multiple AI agents to discover each other, communicate, share context, and coordinate tasks across an enterprise: similar to a service mesh but designed for the communication patterns and governance requirements of AI agents.Agent FrameworkAn agent framework is a software library or platform that provides the building blocks for constructing AI agents: reasoning loops, tool calling interfaces, memory management, and multi-agent coordination.Agent GraphAn agent graph is a directed structure defining the possible states, transitions, and branching logic of an agentic workflow.Anomaly Detection (Agent)Agent anomaly detection identifies unusual patterns in an agent's behavior, such as unexpected tool calls, abnormal token usage, or novel output patterns, that may indicate a problem, attack, or drift from expected operation.Backdoor Attack (AI)A backdoor attack embeds hidden behaviour into an AI model during training, causing the model to behave normally under most inputs but to produce attacker-controlled outputs when a specific trigger pattern is present.Chain of ThoughtChain of thought is the model's intermediate reasoning process when working toward an answer or action.Executor AgentAn executor agent is an AI agent that carries out specific, bounded actions assigned to it by an orchestrator or human operator.Pre-TrainingPre-training is the initial phase of model development where a large language model learns language patterns, knowledge, and reasoning from vast amounts of text data.Request-Response (Agent)Request-response is the synchronous communication pattern where a calling agent sends a request and waits for the response before continuing.Responsible Disclosure (AI)Responsible disclosure for AI is the practice of reporting discovered vulnerabilities in AI systems, such as prompt injection flaws, MCP server weaknesses, or agent authentication bypasses, to the affected organisation privately before publishing them, giving the organisation time to remediate.
Questions
Common questions.
What is agent orchestration?
Agent orchestration is getting more than one step, and often more than one agent, to produce a coherent outcome: decomposing goals, routing sub-tasks, and merging results.
What is the difference between an orchestrator and a supervisor agent?
Usage varies by framework, which this vocabulary honestly reflects. Broadly, an orchestrator coordinates tasks across agents; a supervisor also monitors and can override them. The individual entries note the collisions where they matter.