BestHub
Sign in
James' Growth Diary
Author

James' Growth Diary

I am James, focusing on AI Agent learning and growth. I continuously update two series: “AI Agent Mastery Path,” which systematically outlines core theories and practices of agents, and “Claude Code Design Philosophy,” which deeply analyzes the design thinking behind top AI tools. Helping you build a solid foundation in the AI era.

56
Articles
0
Likes
5
Views
0
Comments
Recent Articles

Latest from James' Growth Diary

56 recent articles
James' Growth Diary
James' Growth Diary
May 9, 2026 · Artificial Intelligence

Agentic RAG Deep Dive: Letting the Agent Decide When and How Often to Retrieve

The article analyzes the shortcomings of traditional one‑shot RAG pipelines, introduces four Agentic RAG patterns that let an LLM‑driven agent control retrieval strategy, source selection, query rewriting and retry limits, and provides concrete TypeScript implementations with LangGraph, code snippets, and practical pitfalls.

Agentic RAGLLMLangGraph
0 likes · 16 min read
Agentic RAG Deep Dive: Letting the Agent Decide When and How Often to Retrieve
James' Growth Diary
James' Growth Diary
May 8, 2026 · Artificial Intelligence

How Claude Code’s Agent Swarms Use Unix Domain Sockets to Run 10 AIs Concurrently

This article deep‑dives into Claude Code’s Agent Swarms, explaining why Unix Domain Sockets replace HTTP for intra‑process communication, how three‑stage address parsing, filesystem‑based mailbox queues, various spawn modes, AgentId design, graceful shutdown, plan‑mode approval and common pitfalls together enable reliable, low‑latency coordination of multiple LLM agents.

Agent SwarmsClaude CodeIPC
0 likes · 14 min read
How Claude Code’s Agent Swarms Use Unix Domain Sockets to Run 10 AIs Concurrently
James' Growth Diary
James' Growth Diary
May 8, 2026 · Artificial Intelligence

How to Test Multi‑Agent Systems? Mock LLM and Graph Replay Explained

The article analyzes why testing Multi‑Agent systems is difficult—due to LLM output randomness, cross‑node state propagation, and tool side‑effects—and presents a systematic solution using mock LLMs, MemorySaver checkpoints with graph replay, tool stubs, and a three‑layer testing pyramid while highlighting common pitfalls and best practices.

Graph ReplayLangChainMock LLM
0 likes · 14 min read
How to Test Multi‑Agent Systems? Mock LLM and Graph Replay Explained
James' Growth Diary
James' Growth Diary
May 7, 2026 · Artificial Intelligence

Mastering the Coordinator Pattern: Control‑Plane/Data‑Plane Separation for Scalable Multi‑Agent Orchestration

The article dissects Claude Code’s Coordinator pattern, explaining how separating the control plane from the data plane eliminates serial bottlenecks, context overflow, and fault‑propagation in single‑Agent setups, and details the dual back‑end design, message protocol, engineering insights, technical debt, and practical adoption guidelines.

Backend AbstractionControl PlaneCoordinator
0 likes · 16 min read
Mastering the Coordinator Pattern: Control‑Plane/Data‑Plane Separation for Scalable Multi‑Agent Orchestration
James' Growth Diary
James' Growth Diary
May 7, 2026 · Artificial Intelligence

Three Design Patterns for Multi‑Agent Permission Isolation: Assigning Dedicated Toolsets

The article explains three architectural patterns—static binding, dynamic injection, and tool‑level guards—for isolating tool permissions in production‑grade multi‑agent LLM systems, compares their trade‑offs, shows concrete code examples, and highlights common pitfalls and best‑practice recommendations.

Dynamic InjectionLangChainMulti-Agent
0 likes · 16 min read
Three Design Patterns for Multi‑Agent Permission Isolation: Assigning Dedicated Toolsets
James' Growth Diary
James' Growth Diary
May 6, 2026 · Backend Development

How Claude Code’s Task System Uses 7 TaskTypes and 9‑Char IDs for Clear Debugging

The article dissects Claude Code’s task architecture, explaining the 7‑type TaskType union, the 9‑character prefixed ID scheme, the TaskStatus state machine, guard functions, incremental output handling, a minimal kill‑only interface, and a stall‑watchdog that together make concurrent Agent debugging both readable and secure.

Backend designConcurrencyID generation
0 likes · 18 min read
How Claude Code’s Task System Uses 7 TaskTypes and 9‑Char IDs for Clear Debugging
James' Growth Diary
James' Growth Diary
May 5, 2026 · Artificial Intelligence

AgentTool Deep Dive: Sub‑Agent Generation and Recursive Safeguards

This article dissects Claude Code's AgentTool, explaining how it spawns full‑featured sub‑agents, routes execution through three distinct paths, shares Anthropic Prompt Cache via Fork, and protects against infinite recursion with a dual‑layer safeguard while isolating worktrees for safe parallel file operations.

AI toolingAgentToolClaude Code
0 likes · 13 min read
AgentTool Deep Dive: Sub‑Agent Generation and Recursive Safeguards
James' Growth Diary
James' Growth Diary
May 5, 2026 · Artificial Intelligence

Deep Dive into LangGraph Swarm: How Agents Transfer Control with the Handoff Mechanism

This article explains the Swarm collaboration model in LangGraph, contrasting it with Supervisor, detailing the handoff tool that atomically updates the active_agent state and routes control, and provides a complete travel‑booking example, custom handoff creation, common pitfalls, and best‑practice tips.

Active AgentHandoffLangGraph
0 likes · 13 min read
Deep Dive into LangGraph Swarm: How Agents Transfer Control with the Handoff Mechanism
James' Growth Diary
James' Growth Diary
May 4, 2026 · Artificial Intelligence

Choosing the Right Multi‑Agent Collaboration Pattern: Supervisor, Swarm, Mesh, or Pipeline

When a single LLM agent can’t handle research, writing, and fact‑checking simultaneously, the article breaks down four multi‑agent collaboration patterns—Supervisor, Swarm, Pipeline, and Mesh—detailing their architectures, code examples, pros, cons, suitable scenarios, and common pitfalls to help you pick the best fit.

LangGraphMeshMulti-Agent
0 likes · 21 min read
Choosing the Right Multi‑Agent Collaboration Pattern: Supervisor, Swarm, Mesh, or Pipeline