مرکزی مواد پر جائیں

Part 2: Agent Workflow Primitives

You understand the paradigm. You've seen what an AI Employee can do (Chapter 56).

Now it's time to build the skills to create your own.

This part is a progressive skills lab. It opens with the Seven Principles of General Agent Problem Solving (Chapter 18) — the operational discipline that turns AI collaboration from random luck into reliable engineering — and then puts each principle to work across file processing, computation, structured data, Linux, version control, and a capstone AI Employee build.

What You'll Learn in Part 2​

  • Chapter 18: The Seven Principles of General Agent Problem Solving — the operational patterns underneath every reliable agent workflow
  • Chapter 19: File Processing Workflows
  • Chapter 20: Computation and Data Extraction
  • Chapter 21: Structured Data and Persistent Storage (with companion lessons 21A on reading SQL and 21B on Postgres as system-of-record)
  • Chapter 22: Linux Mastery for AI Agents
  • Chapter 23: Version Control
  • Chapter 24: Build Your First AI Employee

The Progressive Arc​

Chapter 18 establishes the principles. Each chapter that follows solves a real-world problem while explicitly applying those principles.

Two Interfaces, One Agent​

Throughout this part, you'll use your General Agent through two interfaces:

  • Code (Claude Code): For precise operations: file processing, data analysis, version control, automation scripts
  • Cowork (Claude Desktop): For iterative work: research synthesis, document generation, planning strategies

Each chapter specifies which interface suits the problem. By the end of this part, you'll fluidly switch between both.

A Note on Mindset​

The most important thing you'll develop in this part isn't tool proficiency; it's the ability to decompose real problems into agent-solvable tasks.

Each chapter starts with a problem you recognize from your own work. The solution always follows the same pattern:

  1. Describe what you need (not how to do it)
  2. Verify the agent understood correctly
  3. Execute in small, reversible steps
  4. Persist the output for future reuse

This is the General Agent workflow. By the Project lesson, it will be second nature.