Tier 2 - Mastery: AI-Augmented Development
Tier 2 - Mastery: AI-Augmented Development
Section titled “Tier 2 - Mastery: AI-Augmented Development”Overview
Section titled “Overview”Tier 2 deepens your Claude Code fluency from foundational usage to professional practices. You’ve spent 1-2 weeks shipping real code with Claude. Now learn how to design tools agents actually use, orchestrate complex workflows, secure AI-generated code, review PRs effectively, and scale work across agent teams.
This tier is structured for professional developers and DevOps engineers who completed Tier 1. Product owners can optionally join M06 (Tool Design) and M08 (Security).
Duration: ~7 hours total (5 modules × 90 minutes each) Format: Pre-work reading → Self-directed workshop → Concrete takeaway Prerequisites: Tier 1 completion + 1-2 weeks active Claude Code usage
Module Breakdown
Section titled “Module Breakdown”M06: Tool Design — Building What Agents Can Actually Use
Section titled “M06: Tool Design — Building What Agents Can Actually Use”90 minutes | Intermediate
Learn why mechanical API-to-tool conversion fails and how to design purpose-built MCP tools for agent cognition. Build a custom TypeScript MCP server, consolidate related functions into coherent interfaces, and return agent-efficient formats. Walk away with a working MCP server and reusable design principles.
Audience: Developers, DevOps, optionally Product Owners Hands-on: Custom MCP server in TypeScript
M07: Advanced Workflows — Skills, Subagents, and Hooks
Section titled “M07: Advanced Workflows — Skills, Subagents, and Hooks”90 minutes | Intermediate
Master the workflow composition stack: skills (reusable instructions), subagents (isolated workers), and hooks (deterministic lifecycle scripts). Learn three skill patterns and when to use each. Design a custom security-reviewing subagent. Wire everything into a Feature Development Workflow.
Audience: Developers, DevOps Hands-on: Team skill + custom subagent + hook
M08: Security in the Age of AI-Generated Code
Section titled “M08: Security in the Age of AI-Generated Code”90 minutes | Intermediate
AI accelerates both development and vulnerability creation. Study prompt injection attacks, OWASP Top Ten for LLM Applications, and context rot as a security vector. Build a security-reviewer subagent with concrete checks and integrate it into pre-commit hooks.
Audience: Developers, DevOps, optionally Product Owners Hands-on: Security review workflow + pre-commit integration
M09: AI-Assisted Code Review
Section titled “M09: AI-Assisted Code Review”90 minutes | Intermediate
Code review matters more with AI-generated code, not less. Learn why style is automatable but design judgment is human. Build a /review skill enforcing your team’s checklist. Practice the Writer/Reviewer pattern: one subagent writes, another reviews from fresh context.
Audience: Developers, DevOps Hands-on: /review skill + Writer/Reviewer pattern
M10: Agent Teams and Parallel Orchestration
Section titled “M10: Agent Teams and Parallel Orchestration”90 minutes | Advanced
Scale from subagents to coordinated Agent Teams. Understand when teams add value (large refactors, parallel reviews) vs when simpler patterns work better. Enable experimental Agent Teams, run a real multi-component task, observe coordination and messaging.
Audience: Developers, DevOps Hands-on: Agent Teams on multi-component task + cost analysis
How to Use This Tier
Section titled “How to Use This Tier”- Sequential learning: Modules build on each other. Complete M06 before M07, etc.
- Pre-work first: Spend 15-20 minutes on reading before each workshop.
- Hands-on is required: The takeaway artifacts (skills, hooks, security workflows) become your team’s infrastructure.
- Iterate: After the workshop, refine your implementations in real PRs over the following week.
What You’ll Build
Section titled “What You’ll Build”By the end of Tier 2, you’ll own:
- A custom MCP server for your team’s internal tools
- A team skill following established conventions
- A security-reviewer subagent
- A code review workflow
- Judgment about when Agent Teams pay off
Tier 2 at a Glance
Section titled “Tier 2 at a Glance”| Module | Title | Duration | Audience |
|---|---|---|---|
| M06 | Tool Design | 90 min | Dev, DevOps, PO* |
| M07 | Advanced Workflows | 90 min | Dev, DevOps |
| M08 | Security | 90 min | Dev, DevOps, PO* |
| M09 | Code Review | 90 min | Dev, DevOps |
| M10 | Agent Teams | 90 min | Dev, DevOps |
*PO = Product Owner (optional)
Prerequisites
Section titled “Prerequisites”- ✓ Tier 1 completion (Claude Code basics, MCP fundamentals, hands-on tool creation)
- ✓ 1-2 weeks of active Claude Code usage in real projects
- ✓ Familiarity with TypeScript (for M06)
- ✓ Comfort with bash/shell scripting (for M07, M08, M09)
- ✓ Access to a team Git repository for workshop exercises
Support
Section titled “Support”Each module includes:
- Curated reading list with external links
- Step-by-step workshop guide
- Runnable code examples
- Troubleshooting section
If you get stuck, revisit the “Key Concepts” section or review the References.