Builds
Everything I build is production software with tests, not prototypes or slide decks. Each product implements one or more Design Imperatives. Some are deployed to Vercel, some run locally as MCP servers, some are research prototypes feeding the next paper.
AgentOS + Governance Engine
Patent PendingConstitutional agent OS where governance is structural, not bolted on -- 76 tests, 8 packages, patent-protected.
Simiya (Platform + SimiyaID + Discovery Hub)
Patent PendingOntological discovery platform that finds what must exist, not what already does -- 264 tests, 3 products, swarm architecture.
FORGE
Pipeline app builder that makes AI workflows repeatable, not artisanal -- Glyph Design System, Next.js, Vercel-deployed.
Proforma Intelligence
Patent PendingHybrid financial diagnostic where deterministic computation constrains what AI can say -- Steps 2-3 zero AI calls.
ATLAS
Methodology delivery wizard where consulting IP is the application, not a slide deck -- 6 stages, Fractal Canvas, PDF export.
Practice OS
MCP server that makes consulting knowledge compound instead of evaporate -- 38+ tools, 253-entity graph, 112 tests.
Self-Learning Agent
Patent PendingFour-loop autonomous learning architecture where agents grow capabilities without modifying core axioms -- Python, soul seed governance.
Companion Daemon
Always-on background agent that detects drift when it happens, not when you ask -- continuous practice intelligence.
Intent Architecture Studio
Memory-native content architecture tool that starts from business structure, not keywords -- 7-step analysis, Supabase-backed.
AgentOS MCP Bridge
Scope multiplexer that presents 13 internal servers as one endpoint -- the architectural prerequisite for AgentOS distribution.