Recommended
Claude Code source reading
Start with the series that follows Claude Code through architecture, context, tools, MCP, skills, and multi-agent collaboration.
Open the seriesLien Jack / Agent Learning System
learn agent
Lien Jack writes about agent systems, Claude Code, LLM engineering, and full-stack product engineering.
The archive turns Claude Code runtime, agent harnesses, tool use, SQL, Redis, and product engineering into reusable learning paths.
Core routeAgentFrom prompt loops to controlled execution
Frontier entryClaude CodeSource, tools, MCP, skills, collaboration
Engineering baseFull StackSQL, Redis, services, and interfaces
Agent Runtime
SQL + Redis
Claude Code
Start with these two routes
System route
Build Harness
Learn the control system around agents: execution, permissions, session recovery, observation logs, and verification criteria.
Open Harness
AI Native
An agent is a runtime, not just a chat box
Follow ReAct, context governance, tool calls, permission boundaries, session recovery, and multi-agent collaboration.
- Agent Harness
- Tool Runtime
- MCP / Skill
- Claude Code
Full Stack Core
Learn the database, cache, and service base
SQL, Redis, APIs, queues, and frontend state become one path from data to user experience.
- SQL Modeling
- Redis Cache
- API Design
- Frontend Systems
Product Craft
Turn hard technology into usable product work
Interface structure, motion, and long-form writing make complex engineering learnable and reusable.
- UX Flow
- Three.js
- Framer Motion
- Technical Writing
Built and verified around this stack
View all resourcesAgentClaude CodeMCPSkillSQLRedisReactAstroThree.js
What this notebook keeps building toward
- Turn scattered prompts, experiments, and notes into reusable article series
- Document concrete agent workflow cases instead of stopping at definitions
- Keep refining the site language, typography, and components until the whole place feels intentional
- Build a steady long-form writing rhythm around technical thinking