The Agents SDK separates validation at the input and output boundaries of an agent workflow. Input guardrails evaluate the initial user input, while output guardrails evaluate the final agent output before that result is accepted and returned. This makes B the intended architectural answer. A technical nuance is that current SDK input guardrails support both blocking and parallel execution: with blocking execution, validation completes before agent execution starts; with the default parallel mode, the guardrail can execute concurrently with the agent. Output guardrails, however, operate on the completed final output and always execute after the agent finishes producing it. Guardrails are runtime controls rather than decisions the LLM must explicitly request. OCI's agentic architecture similarly emphasizes governed model-and-tool workflows, making these validation boundaries important when implementing production AI agents. OpenAI GitHub

===============

An embedding is a numerical vector representation of data created by an embedding model, normally implemented using a neural network. Its purpose is to encode semantic characteristics so that items with related meanings are positioned near each other in a multidimensional vector space. Instead of matching only literal keywords, a semantic-search system converts documents and queries into vectors and compares their relative distances or similarities.

Oracle AI Vector Search documentation explains that vector embeddings are mathematical representations describing semantic meaning for content such as text, documents, images, or audio. Oracle further states that modern embeddings are created through neural networks, commonly transformer-based models, although other neural architectures can also be used. This allows Oracle AI Database to store those embeddings using its VECTOR data type and perform similarity searches against them.

A trigger is procedural database logic, a SQL JOIN combines relational data, and a compressed video format is unrelated to semantic representation. Consequently, B is the only technically valid definition. The uploaded question set confirms the same answer.

Study Guide reference/topic: Agentic AI for Oracle AI Database — Oracle AI Vector Search, vector embeddings, semantic similarity, and neural embedding models.

===============

The distinction concerns ownership of the conversation and orchestration flow , not synchronous versus asynchronous execution. In the OpenAI Agents SDK's Manager pattern—also called agents-as-tools—a central manager remains the active user-facing agent. It invokes specialist agents as tools, receives their outputs, synthesizes them, and retains responsibility for the final response. The specialist supports the manager without taking ownership of the conversation.

A Handoff works differently. When the current agent hands the task to another agent, the selected specialist becomes the active agent and takes over the conversation for the remainder of that portion of the run. OpenAI's official Agents SDK documentation explicitly describes the Manager pattern as retaining control and Handoffs as transferring control to a specialized agent.

This distinction allows architects to select centralized orchestration when a single agent must aggregate results or apply common controls, and decentralized handoffs when specialists should directly own particular interactions. Therefore, D states the relationship correctly and matches the uploaded question's answer key.

Study Guide reference/topic: OpenAI Responses API and Agents SDK — multi-agent orchestration, Agents as Tools/Manager pattern, and Handoffs.

===============

Short-term memory compaction is a mechanism for reducing an expanding conversation history into a smaller retained representation while preserving the important information needed for subsequent turns. The uploaded source characterizes this as a summarization process for long conversations , making B the intended answer.

Oracle's current OCI Generative AI documentation states that when conversation compaction is enabled, earlier chat history is automatically condensed as a conversation grows. The purpose is to retain relevant context while lowering token usage and reducing latency. The application can continue using the same conversation ID without manually rebuilding the condensed history.

Conceptually, compaction prevents long-running conversations from continually accumulating every earlier turn verbatim. Instead, previous material is compressed into a more concise memory representation that can still inform future model calls. This is a context-management feature rather than a security masking mechanism.

It also has nothing to do with optimizing Python tool execution or improving network routing. Those belong to separate runtime and infrastructure concerns.

Therefore, B is correct.

Study Guide reference/topic: OCI Enterprise AI Agents — Conversations API, short-term memory, conversation compaction, context retention, token optimization, and latency management.

===============

Oracle AI Database Private Agent Factory is a no-code platform for building, testing, and deploying intelligent AI agents . The uploaded question set marks D as correct, and current Oracle documentation independently confirms that definition.

Oracle describes Private Agent Factory as a platform intended for both business users and engineers. It provides an Agent Builder with visual and drag-and-drop capabilities, enabling users to construct intelligent assistants and workflows without writing conventional application code. The platform can combine pre-built agents, custom agents, reusable templates, enterprise data, LLMs, APIs, databases, and external tools.

The strategic purpose is to lower the engineering barrier for enterprise agent creation while retaining governance and integration with Oracle AI Database capabilities. Current releases include pre-built agents and workflow automation functionality for rapidly creating business-oriented agentic solutions.

It is not an embedding backup product, dedicated Kubernetes deployment manager, or physical training appliance. Those alternatives describe unrelated infrastructure or administration capabilities.

Therefore, D is directly supported by Oracle documentation.

Study Guide reference/topic: Agentic AI for Oracle AI Database — Private Agent Factory, no-code Agent Builder, pre-built agents, custom agents, workflows, and enterprise integration.

===============

Streamable HTTP is the standard MCP transport intended for remote or network-accessible client-server communication. Current MCP architecture documentation distinguishes it from STDIO by explaining that Streamable HTTP uses HTTP POST for client-to-server communication and can optionally use Server-Sent Events for streaming. It enables communication with remote MCP servers and can support standard HTTP authentication mechanisms.

The MCP transport specification further establishes two standard transport mechanisms: stdio and Streamable HTTP . With STDIO, the client launches an MCP server as a local subprocess and exchanges JSON-RPC messages through standard input and standard output. That pattern is therefore most appropriate for local process integration. By comparison, a Streamable HTTP server operates as an independent service and can handle multiple client connections, making it suitable for centralized or cloud-hosted MCP deployments.

Raw TCP sockets and local Unix pipes are not the standard remote MCP transport defined by the protocol. Therefore, C is correct and matches the supplied source material.

Study Guide reference/topic: Model Context Protocol (MCP) Fundamentals — STDIO versus Streamable HTTP transport and remote MCP deployment.

===============

Modern LLM-based agents are specifically designed to avoid requiring every possible execution path to be predetermined. The uploaded course material therefore correctly identifies “Requiring every execution path to be predefined” as the behavior that is NOT characteristic of an agent.

OpenAI defines agents as systems capable of independently accomplishing workflows using an LLM to manage workflow execution and make decisions. An agent can determine when a workflow is complete, correct its actions after receiving observations, and dynamically select tools according to the current state. This differs fundamentally from conventional deterministic automation in which developers encode every branch and execution path beforehand.

Agents commonly pursue objectives across multiple reasoning-and-action cycles. They can invoke external APIs, databases, search systems, or other tools; inspect the resulting observations; and choose subsequent actions. A typical agent loop continues until an exit condition is reached rather than following one permanently fixed sequence.

Predetermined rules may still be used for safety, permissions, and guardrails, but the complete path toward the goal does not need to be pre-scripted.

Therefore, D is the correct answer.

Study Guide reference/topic: Introduction to AI Agents — autonomy, agent loops, observations, dynamic tool use, multi-step goal execution, and deterministic workflows.