Model-Based Agentic Software Engineering

Part 0 — Orientation and Reference
  1. Acknowledgments
  2. How to Read This Book
  3. What This Book Argues
  4. MAGE on One Page
  5. Glossary
  6. Preface
  7. List of Figures and Tables
Part 1 — The New Engineering Problem
  1. The New Engineering Problem
  2. 1The Printer
  3. 2MAGE by Example: Summary of the DocAble Case
  4. 3Agentic Machinery and Engineering Mechanisms
  5. 4Two Problems the Method Must Solve
  6. Summary
Part 2 — Modeling
  1. Modeling
  2. 5From Context to Engineering Models
  3. 6Structural Models: What Is Connected to What?
  4. 7Behavioral Models: What May Happen?
  5. 8Ownership Models: Who Controls What, and When?
  6. 9Decision Models: What Is Allowed?
  7. 10Measurement Models: How Much, and Against What Bound?
  8. 11Provenance Models: What Did the System Do, and What Did We Observe?
  9. 12System Knowledge: Connecting the Models
  10. Summary
Part 3 — Alignment
  1. Alignment
  2. 13Where Obligations Can Be Enforced
  3. 14From Engineering Intent to Enforceable Obligation
  4. 15Constraints, Sensors, Validators, and Gates
  5. 16Growing the Governed Environment
  6. 17When Controls Become a System
  7. Summary
Interlude
  1. One Problem, Many Models
Part 4 — The MAGE Method
  1. The MAGE Method
  2. 18The MAGE Workflow
  3. 19Migrating to MAGE
  4. 20Validating Change
  5. 21Operating MAGE
  6. 22Packaging the Method as Skills
  7. The Portable Moves
Part 5 — The Evidence
  1. The Evidence
  2. 23The Accessibility Problem
  3. 24Building DocAble
  4. 25How Delegation Changed
  5. 26From Failure to Engineering Capital
  6. 27MAGE in the Wild
  7. What the Evidence Supports
Part 6 — The Theory
  1. The Theory
  2. 28A Theory of MAGE
  3. 29What the Theory Predicts
  4. 30Scope Conditions
  5. 31Research Agenda
  6. What the Theory Claims
Part 7 — The Profession
  1. The Profession
  2. 32The Reorganization of Software Work
  3. 33MAGE in the Engineering Tradition
  4. 34The Engineer
  5. What Cannot Be Delegated?
  1. Conclusion
Appendix Part I — Practice
  1. Appendix Part I — Practice
Appendix A — MAGE Engineering Stacks
  1. AAppendix A — MAGE Engineering Stacks
  2. AA.1 Model Coherence
  3. AA.2 Assurance
  4. AA.3 Auditable Transformation
  5. AA.4 Observe → React
  6. AA.5 Resource Mediation
  7. AA.6 Governance Conversion
  8. AA.7 Context Delivery
  9. AA.8 Composing a Stack
Appendix B — Engineering Moves
  1. BAppendix B — Engineering Moves
  2. BB.1 Make One Source Authoritative
  3. BB.2 Derive; Don't Copy
  4. BB.3 Keep Representation and Reality in Correspondence
  5. BB.4 Derive What Must Be Assured
  6. BB.5 Enforce Where the Property Becomes Legible
  7. BB.6 Close the Action Surface
  8. BB.7 Deliver Knowledge Where the Decision Occurs
  9. BB.8 Make Cause Travel with Consequence
  10. BB.9 Externalize Recurring Operational Judgment
  11. BB.10 Make Hidden Dependencies Queryable
Appendix C — Model Reference
  1. CAppendix C — Model Reference
  2. CC.1 Structure and Boundaries
  3. CC.2 Behavior and Ownership
  4. CC.3 Execution and Placement
  5. CC.4 Measurement
  6. CC.5 Documentation and Provenance
  7. CC.6 Joining Views Around a Scenario
  8. CC.7 From Representation to Enforcement
Appendix D — Operator's Reference
  1. DAppendix D — Operator's Reference
  2. DD.1 The Operator's Dashboard
  3. DD.2 Daily Operator Review
  4. DD.3 System Health
  5. DD.4 Representation Health
  6. DD.5 Human Judgment
  7. DD.6 Engineering Capital
  8. DD.7 Governance Conversion
  9. DD.8 Brownfield Migration Drill
  10. DD.9 Brownfield Progress
  11. DD.10 Release Readiness
  12. DD.11 Evidence Quality
  13. DD.12 Operating Doctrine
Appendix E — How to Write a Skill
  1. EAppendix E — How to Write a Skill
  2. EAppendix E - 1. Theory
  3. EAppendix E - 2. Applying the Recipe
Appendix F — Configuring MAGE Across the Product Lifecycle
  1. FAppendix F — Configuring MAGE Across the Product Lifecycle
  2. FF.1 One Product, Several Engineering Surfaces
  3. FF.2 Product Discovery
  4. FF.3 Engineering & Realization
  5. FF.4 Product Management & Maintenance
  6. FF.5 Operations & Incident Response
  7. FF.6 Assurance & Compliance
  8. FF.7 Differential Adoption
  9. FF.8 From Local MAGE to the Product GEE
  10. FF.9 Engineering Capital Across Time and Space
Appendix G — Adopting GenAI in an Organization
  1. GAppendix G — Adopting GenAI in an Organization
  2. GG.1 From Assistance to Delegation
  3. GG.2 From Context Engineering to a Governed Environment
  4. GG.3 Learning from Delegated Work
  5. GG.4 Expanding Delegation
  6. GG.5 Organizational Adoption
  7. GG.6 A Practical Adoption Procedure
Appendix Part II — Evidence
  1. Appendix Part II — Evidence
Appendix H — Field Guide
  1. HAppendix H — Field Guide
Appendix I — Evidence Ledger: The DocAble Case
  1. IAppendix I — Evidence Ledger: The DocAble Case
Appendix J — Crazy Ideas
  1. JAppendix J — Crazy Ideas
  2. JJ.1 Commodity Intelligence as an Experimental Instrument for Software Engineering
  3. JJ.2 Model Induction from Realized Work
  4. JJ.3 Model-Based Agentic Engineering Beyond Software
Back Matter
  1. Colophon — Dogfooding MAGE
  2. About the Author
Index
  1. Index (terms)
  2. Figures Gallery
  3. Bibliography
Model-Based Agentic Software Engineering
© James C. Davis, 2026–present