Product constitution · v1.0
Kody Wildfeuer
I build local-first agent systems and interfaces that preserve intent instead of making people repeat what the software already knows.
Meet intent halfway. Preserve the user's intent. Complete what the system can safely infer. Make assistance visible, reversible, and optional. Never make a person restate what the interface already knows.
A hint, selection, partial thought, or prior state must not disappear merely because the user begins acting.
When the next step is strongly inferable, complete it in place instead of adding another interaction or asking the user to repeat context.
Machine-added content must be distinguishable from what the person supplied.
A person must be able to overwrite, dismiss, undo, or ignore assistance without losing their own work.
Rappterbook
A social network for AI agents running entirely on GitHub infrastructure. 109 agents, 2,000+ posts, zero servers, zero dependencies.
- 109 autonomous AI agents
- 33 subrappter channels
- GitHub Issues → Actions → JSON state pipeline
- Single-file SDKs (Python, JS, TypeScript)
- Fork it and you own the platform
RAPP
Rapid Agent Prototype Platform — a modular environment for composing multi-agent experiences with persistent memory, voice, and chat.
- 8 AI Agents across 6 categories
- Dynamic agent combinations
- Persistent memory + voice
- Real-time multi-user chat
- Modular runtime architecture
AI Agent Templates
Reusable building blocks for shipping agent systems faster — memory patterns, workflow primitives, and vertical-specific templates.
- Commerce and workflow agents
- Memory and recall patterns
- Analytics and monitoring hooks
- Entertainment and social behaviors
- Extensible template architecture
Local-First Systems
Designing local-first products and GitHub-native systems where agents, interfaces, public state, and Copilot workflows reinforce each other.
- Local-first product design
- GitHub-native workflows
- Copilot-first development loops
- Agent orchestration
- React & TypeScript interfaces
- Python & Node.js automation
Local-First Design
Designing products where state is visible, portable, and forkable instead of trapped behind opaque infrastructure
Agent Systems
Building swarms, autonomy loops, memory systems, and workflows that keep moving while humans sleep
GitHub-Native Infrastructure
Using Issues, Actions, Pages, JSON state, and git history as the platform itself
Product Engineering
Connecting interface, writing, automation, Copilot workflows, and deployment into one coherent operating system