The Question That Reveals Everything
Open the system prompt of any AI agent you've built or used. Read it. Now answer this question: if you removed the agent's name and domain, could you tell which agent it was just from how it thinks?
For most agents, the answer is no. And that's the problem.
Most agent prompts are job descriptions. "You are a helpful assistant that does X. Be polite. Be accurate. Follow these rules." Swap the domain, and the agent is interchangeable. It has no cognitive identity, no consistent way of approaching problems, no personality that persists across different types of inputs. It does what you told it to do. It does not think the way you need it to think.
This module teaches you the difference between a job description and an identity. A job description tells an agent what to do. An identity tells an agent how to think. That distinction changes everything about how the agent behaves in situations you didn't anticipate, and those situations are where identity earns its value.