Publications from working systems
Books built from running code, not outlines.
These aren't theoretical manuscripts. Each book is derived from an operational system — the same tools, specs, and pipelines that run the code also assemble the publication. Reproducible, verifiable, and continuously updated.
- 2 publications
- Repository-backed — every claim is testable
- Self-generating pipeline
Executable book · Working repository · Production thesis
MCP with DSPy Theory and Application
An executable book for building spec-driven, policy-enforced MCP agents. Ideas originate in a PKM vault, specs define tool contracts, policy fences govern side effects, and tests enforce guarantees. The result is published as a professional multi-format edition from the same repo that proves the claims.
- Contract-first MCP tooling with JSON-schema validation
- Policy-enforced PKM writes with vault-relative path semantics
- Published in PDF, HTML, EPUB, and manuscript form
- Self-generating: the book teaches the system that built it
Systems engineering · Cognitive infrastructure · Visual manual
The Augmented Mind Forge
A visual engineering manual for designing, operating, and evolving personal cognitive infrastructure with OpenClaw. Covers the full stack: memory systems, model routing, autonomous agents, cron-based automation, and the operator interface between human intention and AI execution.
- Four-system cognitive fusion architecture
- Autonomous agent design patterns and safety fences
- PKM vault as cognitive operating system
- Visual system diagrams and decision flow maps
