Course Outline
The Four-Level Personalisation Stack
Level 1 | Knows – Rules and AGENTS.md
Topics covered:
• Defining project conventions and coding standards
• Documenting architecture and technical constraints
• Creating tool-neutral project guidance
• Maintaining consistency across development teams and AI tools
Level 2 | Can – Skills
Topics covered:
• Creating reusable units of specialist knowledge
• Loading contextual information only when required
• Reducing context size while improving task performance
• Building libraries of reusable workflows and expertise
Level 3 | Reaches – MCP
Topics covered:
• Connecting AI tools to external systems and services
• Accessing repositories, databases and documentation sources
• Extending the capabilities of AI coding assistants
• Implementing secure integrations and governance controls
Level 4 | Acts – Agents
Topics covered:
• Understanding autonomous AI agents and their capabilities
• Reading, writing, testing and revising code autonomously
• Managing goal-driven workflows and delegated tasks
• Establishing oversight and human review mechanisms for agentic systems
Day 1 | Delegation and Extending the Tools
Module 1 | From Assistant to Agent
Topics covered:
• Understanding the difference between AI assistants and autonomous agents
• Inline code completion versus agentic delegation
• How agentic workflows change the way development tasks are structured
• Identifying tasks that are suitable for delegation to agents
• Best practices for collaborating with autonomous AI systems
Module 2 | Delegations That Work Without Babysitting
Topics covered:
• Writing effective instructions for AI agents
• Providing sufficient context and business requirements
• Defining constraints and boundaries for execution
• Establishing clear acceptance criteria and success measures
• Minimising human intervention while maintaining quality
Module 3 | Personalisation Stack and What Applies Where
Topics covered:
• Understanding the four-level personalisation stack
• Using Rules and AGENTS.md to define project conventions
• Determining which personalisation mechanisms apply in different scenarios
• Managing context efficiently across tools and projects
• Creating consistent AI-assisted development environments
Module 4 | Skills and Subagents
Topics covered:
• Creating reusable Skills for common workflows and tasks
• Packaging specialist knowledge for repeated use
• Understanding the role of subagents and isolated contexts
• Delegating bounded tasks to specialised agents
• Improving efficiency through modular AI workflows
Day 2 | Connecting Tools, Parallelism and Governance
Module 5 | MCP: Connect and Build
Topics covered:
• Understanding the principles of the Model Context Protocol (MCP)
• Connecting AI tools to external systems and services
• Integrating browsers, databases, repositories and documentation sources
• Building a custom MCP server
• Managing access control and security considerations
Module 6 | The Disciplined Agentic Workflow
Topics covered:
• Establishing a repeatable AI-assisted development process
• Brainstorming and planning with AI agents
• Building and implementing solutions collaboratively
• Testing and validating generated outputs
• Reviewing and finalising deliverables with appropriate verification steps
Module 7 | Parallel Development
Topics covered:
• Running multiple AI agents simultaneously
• Working with isolated branches and Git worktrees
• Coordinating development activities across parallel workflows
• Merging and validating outputs from multiple agents
• Improving productivity through parallel execution strategies
Module 8 | Risks, Review and Governance
Topics covered:
• Evaluating and vetting external Skills and MCP servers
• Understanding security and governance risks
• Managing permissions and access rights
• Protecting sensitive data and intellectual property
• Establishing review processes and quality assurance practices
Module 9 | AI Adoption in Software Development: Use Cases and Next Steps
Topics covered:
• How organisations are integrating AI into the Software Development Lifecycle (SDLC)
• Real-world use cases and implementation examples from different industries
• Common AI adoption approaches: individual adoption, team-based adoption and organisation-wide enablement
• Typical use cases across the SDLC:
• Requirements gathering and documentation
• Code generation and prototyping
• Testing and quality assurance
• Code review and refactoring
• Documentation and knowledge management
• DevOps and incident management
• Governance models, policies and security considerations
• Measuring productivity and ROI of AI-assisted development
• Building an internal AI adoption roadmap
• Defining practical next steps for participants and their teams
Interactive Discussion Workshop
• Current challenges within the participants' development teams
• Identification of high-value use cases for immediate adoption
• Risks, blockers and organisational considerations
• Creation of an initial action plan for AI integration.
Requirements
Participants should have professional development experience, be comfortable with the terminal, and possess working knowledge of Git. Additionally, they should either regularly use an AI coding tool or have completed the Foundations course.
Audience
The course is tailored for developers already employing AI tools, technical leads responsible for team adoption, and platform or DevOps engineers tasked with building Skills and MCP servers.
Testimonials (2)
Using Claude Code in a more efficient way
Virgil Trif - Frequentis
Course - Claude Code: Agentic AI Development · 1-Day
"I learned the potential of the tool and gained sufficient skills to start using it for my work right away
