Developer Guide

Technical documentation for people who want to understand the implementation details, modify the system, or build custom components.

Note: You don’t need this section to use the system. The main documentation covers everything you need to get set up and productive. This section is for developers, power users, and the technically curious.

What’s Here


  ┌─ System Design ──────────────────────┐  ┌─ Configuration ──────────────────────┐
  │  Architecture Overview               │  │  Context Architecture                │
  │  Components                          │  │  Configuration Files                 │
  │  Data Flow                           │  │  API & Credentials                   │
  │  Scheduling                          │  └──────────────────────────────────────┘
  └──────────────────────────────────────┘
                                            ┌─ Migration ──────────────────────────┐
  ┌─ Building ───────────────────────────┐  │  From OpenClaw                       │
  │  Creating Commands & Skills          │  │  From Scratch                        │
  │  Workflow Mapping (Full)             │  └──────────────────────────────────────┘
  └──────────────────────────────────────┘

Section Guide

System Design

Page	What It Covers	When You Need It
Architecture Overview	Core design principles, system diagram	Understanding how everything fits together
Components	Agents, commands, skill packs — file structures and conventions	Building new agents or commands
Data Flow	Data pipelines, integrations, collection scripts	Connecting new data sources
Scheduling	LaunchAgents, cron, hybrid scheduling patterns	Setting up automated tasks

Context & Configuration

Page	What It Covers	When You Need It
Context Architecture	All 6 context layers, loading mechanisms, refresh cycle	Debugging context loading, optimizing context bundles
Configuration	CLAUDE.md, settings.json, hooks system	Customizing system behavior, setting up hooks
API & Credential Setup	OAuth flows, API keys, 1Password integration	Setting up new integrations

Building Custom Components

Page	What It Covers	When You Need It
Creating Commands & Skills	How to design and build custom thinking tools, commands, and skill packs	Building custom workflows and tools
Workflow Mapping (Full Guide)	Complete 6-step methodology with templates and scoring	Translating a role’s entire playbook into the system

Migration

Page	What It Covers	When You Need It
From OpenClaw	Migrating from a Telegram-based agent system	Coming from OpenClaw/ClaudeClaw
From Scratch	Setting up from zero with a recommended timeline	No existing agent system

Quick Reference

Key File Locations

File	Purpose
`~/.claude/CLAUDE.md`	Global agent instructions
`~/.claude/settings.json`	Permissions, hooks, MCP servers
`~/.claude/commands/`	Slash command definitions
`~/.claude/agents/`	Agent persona definitions
`~/.claude/skills/`	Skill pack directories
`~/Vault/0_InnerContext/`	Context layers and routing
`~/Vault/_System/`	System-level configuration

Architecture Principles

Deterministic + Probabilistic — Scripts for repeatable tasks, LLM for reasoning
Hooks enforce correctness — Pre/post validation on every action
PARA structure — Consistent organization across the vault
Credential vault — No hardcoded secrets
Layered context — Right information at the right time

Component Types

Type	What It Is	File Location
Agent	Specialized persona (.md file)	`~/.claude/agents/`
Command	User-triggered workflow (.md file)	`~/.claude/commands/`
Skill Pack	Multi-step tool (directory with SKILL.md)	`~/.claude/skills/`
Script	Deterministic data pipeline (Python)	`~/.claude/scripts/`
Hook	Pre/post action validator	`~/.claude/hooks/`