Glossary
Precise definitions of the key terms, concepts, and techniques that make up the vocabulary of agent engineering. 13 terms
A
- AgentCore Concepts
- An AI system that autonomously perceives its environment, reasons about goals, takes actions, and adapts based on feedback — typically over multiple turns without explicit step-by-step human direction.
- Agent-to-Agent Protocol (A2A)Protocols & Standards
- An open protocol introduced by Google that defines a standard communication layer for autonomous AI agents to discover, invoke, and collaborate with one another across organizational and framework boundaries.
C
- Chain of ThoughtPrompting & Reasoning
- A prompting technique that instructs a language model to output intermediate reasoning steps before producing a final answer, dramatically improving performance on multi-step problems.
- Context WindowCore Concepts
- The maximum number of tokens a language model can process in a single forward pass — encompassing the system prompt, conversation history, retrieved documents, tool results, and the model's own generated output.
F
- Function CallingCore Concepts
- A native model capability — first popularized by OpenAI — that allows an LLM to emit structured, schema-validated requests to invoke developer-defined functions, replacing ad-hoc text parsing with reliable tool invocation.
H
- Human-in-the-LoopArchitecture
- A design pattern where an autonomous agent pauses at defined checkpoints to request human review, approval, or correction before proceeding — balancing automation speed with human oversight and accountability.
M
- Model Context Protocol (MCP)Protocols & Standards
- An open standard introduced by Anthropic that defines a uniform interface for connecting LLM-based agents to external tools, data sources, and services — replacing ad-hoc integrations with a single, reusable protocol.
- Multi-Agent SystemArchitecture
- An architecture in which multiple specialized AI agents collaborate — through shared memory, message passing, or an orchestrator — to solve tasks too complex or parallel for a single agent.
O
- OrchestrationArchitecture
- The coordination of multiple agent calls, tools, and models to accomplish a complex, multi-step task — including managing state, routing, error handling, and human-in-the-loop checkpoints.
P
- Prompt InjectionSecurity
- An attack in which malicious content in an agent's environment — web pages, documents, tool outputs — overrides the developer's system prompt instructions, hijacking the agent's behavior.
R
- ReActPrompting & Reasoning
- A prompting framework that interleaves reasoning traces and action calls in a single LLM output, enabling agents to think through a problem step-by-step while simultaneously grounding reasoning in real-world observations.
- Retrieval-Augmented Generation (RAG)Memory & Context
- A technique that enhances LLM outputs by retrieving relevant documents or data from an external knowledge source and injecting them into the model's context before generation.
T
- Tool UseCore Concepts
- The capability of a language model to invoke external functions, APIs, or services as part of generating a response, enabling agents to take actions beyond text generation.