Articles
Short, exact articles on Claude, AI engineering and software concepts. Each one goes with a video and links the official docs.
Every article is written from the official documentation and the exam guide, and checked line by line. Start with the CCAR-F study plan if you are preparing the Claude Certified Architect exam.
The Agentic Loop Explained: How AI Agents Decide
The agentic loop explained with Claude: model call, tool call, tool result, repeat; why stop_reason ends the loop, and three wrong ways to stop an agent.
21 October 2026
Claude Agent SDK Tutorial: Build Your First Agent
Claude Agent SDK tutorial: install it, run a first agent with query(), then control tools, permissions, limits, hooks, subagents and sessions.
21 October 2026
Claude Agent SDK vs Claude Code vs Claude API vs MCP
Claude Agent SDK vs Claude Code, the Claude API and MCP: one question tells them apart (who runs the agent loop?), with a table and the exam trap.
21 October 2026
Claude API Pricing Explained: Every Model and Discount
Claude API pricing as of October 2026: input and output prices per model, prompt caching, the 50% batch discount, long context, tool costs and a worked example.
21 October 2026
Claude Certified Architect Exam: CCAR-F Study Plan
The Claude Certified Architect exam (CCAR-F) in one page: 60 questions, 720 to pass, five domains, six scenarios, and a free study plan for each domain.
21 October 2026
Claude Code Best Practices: 10 Habits That Work
Claude Code best practices from the official guide: give Claude a check to run, explore then plan, specific prompts, a short CLAUDE.md, hooks, subagents and /clear.
21 October 2026
Claude Code GitHub Actions: Setup, Reviews and CI
Claude Code GitHub Actions explained: /install-github-app, @claude mentions, an automated PR review workflow, claude -p in CI with JSON output, and keeping costs in check.
21 October 2026
Claude Code Hooks: Events, Examples and Exit Codes
Claude Code hooks explained: what they are, the key events (PreToolUse, PostToolUse, Stop), where to configure them, exit code 2, and hooks vs prompt rules.
21 October 2026
Claude Code Plan Mode: When and How to Use It
Claude Code plan mode explained: how to turn it on (Shift+Tab or /plan), how to review and approve the plan, the permission modes, and when direct execution is better.
21 October 2026
Claude Skills Explained: SKILL.md, Loading and Examples
Claude skills explained: the SKILL.md file, how skills load in three levels, where they work (Claude Code, API, claude.ai), and the frontmatter that matters.
21 October 2026
Claude Subagents: Create and Use Them in Claude Code
Claude subagents explained: their own context window, tools and permissions, the built-in Explore and Plan agents, the .claude/agents file format, and when to delegate.
21 October 2026
CLAUDE.md Guide: What to Put in It and Where It Goes
CLAUDE.md explained: the four levels (managed, user, project, local), how Claude Code loads them, @imports, path-scoped rules, and how to keep the file short.
21 October 2026
Claude API Basics: Messages, max_tokens and stop_reason
Claude API basics: the Messages API request and response, the system prompt, why you resend the history, max_tokens, and what each stop_reason means.
21 October 2026
Claude Structured Output: JSON Schemas That Hold
Claude structured output explained: JSON outputs with output_config.format, strict tool use, the schema rules, and why valid JSON can still be wrong.
21 October 2026
LLM Tool Calling with Claude: How Tool Use Works
LLM tool calling explained with Claude tool use: define a tool, read the tool_use block, run it, return a tool_result. Plus tool_choice and strict mode.
21 October 2026
Few-Shot Prompting: How Many Examples, Which Ones
Few-shot prompting explained with Claude: how many examples to use, how to aim them at the hard cases, the example tags, and the mistakes that teach the wrong pattern.
21 October 2026
LLM Context Window Explained: Claude's Working Memory
The LLM context window explained with Claude: what counts toward it, context rot, 1M-token windows, overflow behavior, and compaction for long agents.
21 October 2026
LLM Hallucination: 7 Ways to Reduce It with Claude
LLM hallucination explained with Claude: let it say I don't know, quote before answering, verify with citations, restrict to your documents, and check in code.
21 October 2026
Multi-Agent Systems with Claude: Coordinator Pattern
Multi-agent system design with Claude: the coordinator and subagents pattern, when it beats a single agent, what it costs in tokens, and how to delegate.
21 October 2026
What Is MCP? The Model Context Protocol Explained
What is MCP? How the Model Context Protocol connects AI apps to tools and data: hosts, clients, MCP servers, tools vs resources, and setup in Claude Code.
21 October 2026