MCP: Integrating Claude with Applications and Tools

1. Course Description

The Advanced Claude Code: Agentic Software Development course provides the knowledge and skills needed to design, implement, and optimize advanced application development processes using artificial intelligence agents, leveraging Claude Code, Anthropic's agentic programming platform.

The training delves into Claude Code's capabilities for planning and executing complex tasks, coordinating specialized agents, automating development processes, integrating external tools, and managing software projects using natural language instructions.

Unlike traditional programming assistance tools, Claude Code enables you to handle complete software lifecycle activities, from requirements analysis and technical planning to code generation, testing, change review, and continuous integration.

During the course, participants will learn to create specialized agents and reusable workflows using Subagents, Agent Teams, Skills, Hooks, and the Model Context Protocol (MCP). Integration with Git repositories, CI/CD pipelines, and enterprise development environments will also be covered.

Special attention will be given to context engineering, security, agent supervision, token consumption optimization, and the application of best practices to maintain the quality and reliability of AI-generated software.

The training combines advanced knowledge, hands-on labs, and an integrative project focused on real-world scenarios.

By the end, each participant will be able to design and execute an agentic development workflow that combines specialized agents, automation, tool integration, and continuous software validation.

2. Intended For

This course is intended for technology professionals who are already familiar with the fundamentals of artificial intelligence-assisted programming and wish to move toward software development models based on specialized agents and automated processes.

It is especially geared toward:

• Experienced software developers and programmers who want to automate complex activities and increase their productivity using AI agents.

• Software and solutions architects interested in designing new agentic development methodologies and architectures.

• Software engineers and technical leads looking to incorporate specialized agents into enterprise projects.

• DevOps and Platform Engineering professionals interested in automating testing, reviews, continuous integration, and deployment processes.

• Tech Leads and development team managers who need to establish best practices for adopting, supervising, and governing agentic tools.

• Technology consultants and artificial intelligence specialists who want to develop advanced solutions using Claude Code and its integration capabilities.

• Automation and internal tool development professionals looking to connect programming agents with services, APIs, and enterprise systems.

• Engineering teams from companies and organizations interested in transforming their development processes using agentic artificial intelligence.

Prerequisites

Experience in software development and knowledge of programming, Git, the terminal, and development environments are recommended.

It is advisable to have completed the Claude Code 101: Artificial Intelligence-Assisted Programming course or have equivalent knowledge.

Familiarity with software testing, APIs, and basic continuous integration concepts is also recommended.

3. Course Objectives

General Objective

To enable participants to design, implement, and manage advanced agentic software development processes using Claude Code by creating and coordinating artificial intelligence agents, integrating external tools, automating workflows, and applying best practices for quality, security, and supervision, with the aim of improving the efficiency and reliability of the application development lifecycle.

Specific Objectives

By the end of the training, participants will be able to:

1. Understand the principles of agentic software development, identifying the differences between assisted programming and executing tasks using supervised autonomous agents.

2. Configure Claude Code for advanced projects, managing instructions, context, permissions, and execution environments.

3. Design planning and execution strategies, using work modes that allow you to tackle complex tasks and multi-component projects.

4. Create and configure specialized subagents, assigning them tasks such as analysis, development, testing, documentation, and code review.

5. Coordinate multiple AI agents, distributing tasks and consolidating results using agentic collaboration mechanisms.

6. Develop Skills and reusable workflows, standardizing development procedures within teams.

7. Implement Hooks and automations, incorporating controls and actions that run during the development cycle.

8. Integrate Claude Code with external tools using MCP, connecting repositories, services, APIs, and other information sources.

9. Automate software engineering activities, including testing, static analysis, code review, and documentation generation.

10. Integrate Claude Code into CI/CD workflows, using version control tools and automation platforms.

11. Apply security and governance mechanisms, defining permissions, operating limits, and human review procedures.

12. Optimize the use of context, models, and resources, reducing unnecessary operations and controlling execution costs.

13. Evaluate the quality and productivity of agentic processes, using technical and operational indicators.

14. Develop an advanced software project, applying specialized agents, integrated tools, and automated testing and delivery.

4. Course Content

Unit 1. Advanced Architecture and Fundamentals of Agentic Development with Claude Code

1.1. The evolution of assisted programming toward agentic software engineering.

1.2. Advanced architecture and operation of Claude Code.

1.3. The lifecycle of an agentic task: planning, execution, observation, and validation.

1.4. Advanced configuration of projects and work environments.

1.5. Managing persistent instructions using CLAUDE.md.

1.6. Context engineering: information selection, session management, and compaction.

1.7. Execution modes, planning, and permission control.

1.8. Strategies for breaking down complex tasks and reducing execution errors.

1.9. Managing parallel sessions and isolating changes using Git Worktrees.

1.10. Best practices for structuring projects prepared for agentic development.

Practice: Configure an existing software project for development with Claude Code, defining persistent instructions, work rules, permissions, and a planning strategy to implement new features.

Unit 2. Creating and Coordinating Specialized Agents

2.1. Concept and architecture of multi-agent systems applied to software development.

2.2. Introduction to Claude Code Subagents.

2.3. Creating and configuring specialized agents.

2.4. Defining responsibilities, instructions, tools, and permissions for each agent.

2.5. Designing agents for requirements analysis and architecture.

2.6. Implementing agents for development, debugging, and testing.

2.7. Configuring code review and security agents.

2.8. Introduction to Agent Teams and coordinating independent sessions.

2.9. Task distribution, parallel execution, and result consolidation.

2.10. Managing dependencies, conflicts, and agent supervision.

Practice: Create a team of specialized agents for analysis, development, testing, and code review, assigning each one specific tasks within a common project.

Unit 3. Skills, Hooks, and Advanced Process Automation

3.1. Claude Code extensibility architecture.

3.2. Introduction to Skills and their role in agentic processes.

3.3. Designing and creating custom Skills.

3.4. Structuring reusable instructions, resources, and procedures.

3.5. Developing specialized workflows for routine tasks.

3.6. Concept and operation of Hooks.

3.7. Configuring Hooks associated with execution cycle events.

3.8. Automating validations, formatting, and static code analysis.

3.9. Introduction to plugins and distributing reusable configurations.

3.10. Applying preventive controls and error recovery procedures.

Practice: Create a Skill for code review and implement Hooks that run automatic checks after modifying files and block unauthorized operations.

Unit 4. Integrating Claude Code via Model Context Protocol (MCP)

4.1. Fundamentals and architecture of Model Context Protocol.

4.2. The role of MCP in agentic development systems.

4.3. Client-server architecture and capabilities of MCP servers.

4.4. Installing and configuring MCP connections in Claude Code.

4.5. Integration with project management tools and code repositories.

4.6. Connecting to databases, web services, and enterprise APIs.

4.7. Using external tools to expand agent capabilities.

4.8. Authentication, authorization, and credential protection.

4.9. Managing permissions, trust, and access limits.

4.10. Diagnosing, testing, and maintaining MCP integrations.

Practice: Configure Claude Code with a test MCP server to query an external information source and use the results to implement a software feature.

Unit 5. Test Automation, Software Quality, and Security

5.1. Quality assurance strategies in agent-assisted development.

5.2. Automatic generation of unit and integration tests.

5.3. Designing functional tests from requirements.

5.4. Applying test-driven development (TDD) methodologies with Claude Code.

5.5. Detecting and correcting defects using specialized agents.

5.6. Integrating static analysis and code evaluation tools.

5.7. Automating code reviews and vulnerability detection.

5.8. Validating dependencies and external components.

5.9. Error management, recovery, and change reversal.

5.10. Defining acceptance criteria and human supervision of generated code.

Practice: Configure a quality assurance process using Claude Code agents, incorporating test generation, code analysis, issue detection, and automatic validation before integrating changes.

Unit 6. Continuous Integration, DevOps, and Agentic Process Optimization

6.1. Integrating Claude Code into the DevOps lifecycle.

6.2. Advanced Git repository management and branching strategies.

6.3. Automating tasks using Claude Code in non-interactive mode.

6.4. Integrating Claude Code with GitHub Actions.

6.5. Designing continuous integration and continuous delivery (CI/CD) pipelines.

6.6. Automating Pull Request reviews and documentation generation.

6.7. Controlled execution of agents in development and validation environments.

6.8. Managing credentials, permissions, and secrets in automated processes.

6.9. Optimizing context, token consumption, and model selection.

6.10. Monitoring executions, costs, quality, and productivity.

Practice: Implement a continuous integration workflow in which Claude Code assists with change review and test preparation, using GitHub Actions and approval controls.

Unit 7. Practical Project: Developing a Software Solution Using AI Agents

7.1. Identifying an enterprise use case and defining the scope.

7.2. Analyzing functional, technical, and security requirements.

7.3. Designing the application and agentic system architecture.

7.4. Configuring the project, instructions, and development environment.

7.5. Creating specialized agents for different activities.

7.6. Integrating Skills, Hooks, and MCP tools.

7.7. Incremental feature development using coordinated agents.

7.8. Automating testing, code review, and quality control.

7.9. Integrating with Git and configuring a CI/CD pipeline.

7.10. Evaluating performance, documenting, and presenting results.

Final Practice: Develop an enterprise application using an agentic workflow with Claude Code. The participant must implement specialized agents, integrate external tools, automate testing, and establish a controlled change integration procedure.

The project must demonstrate that agents can collaborate across different development phases while maintaining human supervision and validation mechanisms.

5. Learning Outcomes

Upon completing the seven units, participants will be able to:

• Design development environments prepared to execute complex tasks using Claude Code.

• Create and coordinate specialized agents for different software engineering activities.

• Implement Skills and Hooks to automate procedures and quality controls.

• Integrate Claude Code with enterprise services and tools using MCP.

• Automate testing, code reviews, and maintenance processes.

• Incorporate Claude Code into CI/CD pipelines and DevOps workflows.

• Apply security mechanisms, permission controls, and human validation.

• Optimize resource consumption and execution costs.

• Evaluate the impact of agents on development productivity and quality.

• Develop software solutions using a supervised agentic architecture.

6. Methodology

The training is conducted using a practical methodology based on software engineering projects and real-world enterprise development scenarios.

Participants will work with prepared repositories and applications to experiment with agents, automations, and integrations.

The course will combine technical demonstrations, guided labs, problem-solving, and a final integrative project.

The training process will follow four phases:

7. Evaluation and final project

Evaluation will be based on practical activities associated with the units and the development of an integrative project.

Proposed project: Business management platform developed with Claude Code agents

Each participant must design and implement a business management web application that incorporates information querying, recording, and updating functionalities, using AI agents during development.

The project must include:

1. Git repository with Claude Code instructions and configuration.

2. Specialized agents for architecture, programming, testing, and review.

3. At least one reusable Skill for a project activity.

4. Hooks to run automatic technical validations.

5. An MCP integration with a test service or data source.

6. Features developed and verified using agents.

7. Unit and integration tests.

8. CI/CD pipeline with automated validations.

9. Technical documentation of the solution.

10. Results evaluation report and identified improvements.

Evaluation criteria

8. Benefits for companies

Adopting Claude Code in advanced development environments allows organizations to evolve from using individual programming assistants to software engineering processes supported by specialized agents.

Results will depend on the type of project, the available tools, and the level of supervision. The training will enable the identification of appropriate indicators to measure the actual improvement.

9. Resources and technical requirements

To complete the labs, it is recommended to have:

• A Claude account with access to Claude Code or authorized API access.

• A development environment compatible with Claude Code.

• Visual Studio Code or another compatible editor.

• Git and access to a test repository.

• Access to GitHub Actions or an equivalent continuous integration platform.

• A test environment capable of running applications and tests.

• A lab MCP server and authorized integration services.

Practices involving external services must use lab credentials, dummy data, and limited permissions. Some advanced capabilities may depend on the subscribed plan or their availability in the environment.

10. Expected results for the participant