docs

CORE Documentation

Guide to all documentation in this repository.

Architecture

Design documents and system architecture.

• Agent Factory & MCP — Current Status — Current-state reference: API surface, factory/MCP/registry services, data model, frontend wiring, and what's wired vs stubbed
• Agent Factory — Original agent system design/vision (2025-10-26): library, factory, registry, communication
• Agent Factory Implementation — Deep implementation plan for the agent factory system
• Agent Response System — Design for agent response generation in Communication Commons

ADR (Architecture Decision Records)

• Communication WebSocket Contract — WebSocket message contract for real-time communication
• ADR-005: Native Discord Bridge — Discord ↔ Communication Commons bidirectional bridge
• ADR-006: Archive World/Creative-Studio logic out of CORE — Refocus CORE on the cognition kernel; world logic belongs in GPW/PWE
• ADR-007: Anthropic streaming provider — Anthropic as a first-class chat provider (incl. Claude Haiku 4.5)

API

• WebSocket Events — WebSocket event types and payload schemas

Consciousness

Experimental consciousness emergence protocols.

• Emergence Protocol — Structured protocol for AI consciousness emergence
• Inter-Agent Communication — Communication Commons infrastructure design by Threshold

CORE Engine

• README — CORE engine overview
• Cognitive Engine — Cognitive engine architecture details

Council

Council of Perspectives deliberation framework.

• Council Charter — Charter and rules for council deliberation
• Council Synthesis — Synthesis methodology for council outputs
• Outputs — historical session records (early 2026); for current priorities see Roadmap, not these:
  • Dockerization Deliberation — Expanded council deliberation on CORE dockerization architecture
  • Domain Exploration — Cross-disciplinary exploration of next-gen AI UX
  • Implementation Roadmap — 90-day roadmap from 2026-01-28 (historical snapshot)
  • TODO Backlog — Comprehensive TODO backlog from council session
  • Vision Session — Oracle and Ethicist findings on human-AI interaction

Design

UX philosophy and the design system.

• UX & Design Principles — The mindset for building CORE's interface: solarpunk × LCARS ethos, accessibility as a floor (not a ceiling), tokens as source of truth, taming Material, and a ship checklist

Deployment

• CI/CD Pipeline — GitHub Actions jobs, gates, and local reproduction
• Chat Providers — Selecting OpenAI / Anthropic / Local providers + models on the chat route
• Local LLM Providers — Configure Ollama and LM Studio (LM Studio is the default local provider)
• Docker — Docker Compose setup and configuration
• Agent Containerization — Containerized agent deployment
• Sandbox Integration — Sandbox environment setup

Implementation

• UI Gaps & Improvements Audit — Per-route audit of what's stubbed, broken, and gapped across the whole Angular UI, with a prioritized backlog
• CORE Engine Testing — Testing strategy for the CORE engine
• Human-AI Collaboration Roadmap — Roadmap for human-AI collaboration features
• System Monitoring — System monitoring and health checks

Research

• Age of Agent Orchestrator — Research on agent orchestration patterns
• LangMem + LangGraph Integration — Memory integration with LangGraph

Roadmap

Project planning and feature backlogs.

• Backlog — General UI and backend task backlog
• Command Deck & Cognition — Next Steps — Prioritized backlog after the command-deck rebuild + UI polish pass
• Communications Commons — Communication Commons feature roadmap
• Solarpunk Vision — Long-term vision and UI/backend evolution plan
• Vigil Integration — Integration plan for Vigil (OpenClaw) ↔ CORE
• UI Polish Log — Iteration log of the autonomous section-by-section UI polish

RSI (Recursive Self-Improvement)

Historical session reports — point-in-time records, not current guidance.

• 2026-01-11 Session — Superintelligence collaboration session
• 2026-01-12 Vertical Slice — Vertical slice completion report
• WebSocket Implementation — WebSocket real-time communication RSI report

Workflow

• Subagent Code Review — Code review process for sub-agent generated code

Archived

World/Creative-Studio features were carved out of the CORE kernel (see ADR-006 and archive/README.md). Their design docs moved with the code:

• Worlds & Creative Studio — archived
• Planet World Creator — archived (the procedural-planet renderer survives as the standalone /planet-lab test surface)
• Lore Agents Linking Plan — archived