AI Ecosystem CourseLearn with Q
AI Ecosystem · In Development

Agents, MCP, And Tool Use

See how AI agents choose steps and use tools to complete a task.

About This Course

Introduction

An assistant becomes more consequential when it can choose steps and use tools. Before calling it an agent, ask what it can observe, decide, and change. What are its limits? How would we know it succeeded? Answer those questions using evidence about the specific system.

Building on application architecture in AI Application Architecture and decision authority in Decisions, Forecasting, Optimization, And Causality, we will examine workflows, planning loops, tools, state, orchestration, MCP, human review, evaluation, and recovery. Your project specifies a research agent that investigates a question, analyzes and compares information, and prepares a briefing. You will also deliberately break it and explain what happened in a postmortem.

Start with control loops, then interfaces and permissions, then operating examples. Use fictional data for the practice. Keep tools read-only apart from creating local drafts, so you can investigate failures without changing real external records.

Planned Lessons

  1. Workflow Or Agent?Choose between fixed steps and an AI system that decides what to do next.
  2. The Bounded Control LoopFollow an AI agent as it checks the situation, takes an action and looks at what happened.
  3. Planning And ReplanningSee how an agent makes a plan and changes it when a step produces a new result.
  4. Tool Interfaces And ContractsDefine what a tool accepts, what it returns and which actions it can carry out.
  5. MCP's Place In The SystemUnderstand how MCP helps an AI app connect to outside tools and information.
  6. Resources, Prompts, And ToolsSeparate information an agent can read, instructions it can reuse and actions it can take.
  7. Authentication, Authorization, And ConsentLimit an agent’s access and decide when a person must give permission.
  8. State, Checkpoints, And Durable ExecutionSave enough information for an interrupted agent to continue without repeating completed actions.
  9. Orchestration And Dependency ManagementPut dependent tasks in the right order and manage the services needed to carry them out.
  10. Single-Agent And Multi-Agent SystemsCompare giving one agent a task with dividing the work between several agents.
  11. Human-In-The-Loop ReviewPlace human checks before actions where a mistake would matter.
  12. Failure Modes And Compounding ErrorsFollow how an early mistake can affect the later steps an agent takes.
  13. Evaluating Trajectories And OutcomesReview how an agent worked as well as whether it reached the right result.
  14. Agent Economics And Operating LimitsSet limits on an agent’s spending, running time and number of actions.