Mental Models

    Mental models are frameworks for thinking about agentic systems. They're evaluated by utility rather than correctness—a model is valuable when it enables better design decisions, not because it's "true" in some absolute sense.


    Mental Models

    Model Description
    Pit of Success Design systems where the easiest path is also the correct path
    Prompt Maturity Model Seven levels of prompt sophistication from static to meta-cognitive
    Specs as Source Code Specifications are the primary programming surface in agentic systems
    Context as Code Treat agent knowledge like software, not documents
    Execution Topologies A measurement framework for understanding and improving agent execution structures

    Connections

    • To Foundations: The four pillars framework (prompts, models, context, tools) is itself a mental model for decomposing agentic systems
    • To Patterns: Patterns are mental models made concrete through implementation—abstract frameworks translated into specific architectural decisions