Full Walkthrough: Workflow for AI Coding — Matt Pocock

Learn to use Anthropic's Claude Agent SDK (formerly Claude Code SDK) for AI-powered development workflows! https://platform.claude.com/docs/en/agent-sdk/overview https://x.com/trq212 **AI Summary** This workshop by Thariq Shihipar (Anthropic) details the architecture and implementation of the **Claude Agent SDK**. The session moves from high-level theory—defining "agents" as autonomous systems that manage their own context and trajectory—to a live-coding demonstration. Shihipar builds an agent "Harness" from scratch, implementing the core **Agent Loop** (Context Thought Action Observation), integrating the **Bash tool** for general computer use, and demonstrating **Context Engineering** via the file system to maintain state across long tasks. **Timestamps** 00:00 Introduction: Agenda and the "Agent" definition 05:15 The "Harness" concept: Tools, Prompts, and Skills 10:10 Live Coding Setup: Initializing the Agent class and environment 15:45 implementing the "Think" step: Getting the model to reason before acting 25:20 The Agent Loop: connecting `act`, `observe`, and `loop` 33:10 Tool Execution: Handling XML parsing and tool inputs 42:00 The "Bash" Tool: Giving the agent command line access 49:30 Safety & Permissions: "ReadOnly" vs "ReadWrite" file access 58:15 Context Engineering: Using `ls` and `cat` to build dynamic context 01:05:00 The "Monitor": Viewing the agent's thought process in real-time 01:12:45 Handling "Stuck" States: Feedback loops and error correction 01:21:20 Multi-turn Complex Tasks: Building a "Research Agent" demo 01:35:10 Refactoring patterns: "Hooks" and deterministic overrides 01:48:39 Q&A: Reproducibility, helper scripts, and non-determinism 01:50:31 Q&A: Strategies for massive codebases (50M+ lines) 01:52:00 Closing remarks and future SDK roadmap * **Evolution of AI Capabilities:** Shihipar argues we are shifting from **LLM Features** (categorization, single turn) to **Workflows** (structured, multi-step chains like RAG) to **Agents**. He defines agents as systems that *"build their own context, decide their own trajectories, and work very autonomously"* rather than following a rigid pipeline. * **The Claude Agent SDK Architecture:** The SDK is built directly on top of **Claude Code** because Anthropic found they were *"rebuilding the same parts over and over again"* for internal tools. * **The Harness:** A robust agent requires more than just a model; it needs a "Harness" containing Tools, Prompts, a **File System**, Skills, Sub-agents, and Memory. * **Opinionated Design:** The SDK bakes in lessons from deploying Claude Code, specifically the "opinion" that general computer use (Bash) is often superior to bespoke tools. * **The Power of the Bash Tool:** A key technical insight is that the **Bash tool** is often the most powerful tool for an agent. Instead of building custom tools for every action (e.g., a specific API wrapper for a file conversion), giving the agent access to the shell allows it to use existing software (like `ffmpeg`, `grep`, or `git`) to solve problems flexibly, similar to how a human developer works. * **Context Engineering:** Shihipar introduces the concept of **Context Engineering** via the file system. Instead of just "Prompt Engineering," the agent uses the file system to manage its state and context. * **Files as Memory:** The agent can write to files to "remember" things or create its own documentation (e.g., `CLAUDE.md`) to ground future actions. * **Verification:** The file system serves as a ground truth for the agent to verify its work (e.g., checking if a file was actually created). * **The Agent Loop & Intuition:** Building a successful agent loop is described as *"kind of an art or intuition"*. The loop generally follows a **Gather Context Take Action Verify Work** cycle. Shihipar emphasizes that this loop allows the agent to self-correct, a capability missing from rigid workflows. * **Strategies for Determinism (Hooks):** During the Q&A, a technique for controlling agent behavior is discussed: **Hooks**. * If an agent hallucinates or skips a step (e.g., guessing a Pokemon stat instead of checking a script), a hook can intercept the response and inject feedback: *"Please make sure you write a script, please make sure you read this data."* * This enforces rules like "read before you write" without retraining the model. * **Scaling to Large Codebases:** For massive codebases (50M+ lines), standard tools like `grep` or basic context window stuffing fail. * **Semantic Search Limitations:** Shihipar notes that while semantic search is a common solution, it is *"brittle"* because the model isn't trained on the specific semantic index. * **Solution:** He recommends good **"Claude MD"** files (context files) and starting the agent in a specific subdirectory to limit scope, rather than trying to index the entire 50M lines at once.

Visit sintra.ai/leonvanzyl or use my promo code leonvanzyl at checkout to receive limited time offer of 72% off on all plans 🚀 Free resources: source code, prompts, workflows + peer discussion https://skool.com/leonvanzyl 🧪 Agentic Labs: AI coding courses, live Q&A, weekly builder challenges, direct access to me + a serious builder community https://skool.com/agentic-labs OpenCode is the open source coding agent that is quickly becoming a serious alternative to Claude Code. In this opencode tutorial, you will learn how to install OpenCode, connect free AI models through OpenCode Zen or premium providers like OpenAI, and master advanced features including agent skills, sub agents, MCP servers, and background agents. ⏱️ CHAPTERS: 00:00 - OpenCode intro and overview 00:39 - Installing OpenCode in seconds 01:25 - Connecting providers and OpenCode Zen 02:20 - Free AI coding models walkthrough 03:46 - Scaffolding a Next js project 04:55 - Subscriptions vs API keys 06:00 - Connecting OpenAI and GPT 5 07:30 - Installing agent skills from skills sh 10:30 - AGENTS md memory file setup 13:00 - Sub agents and context window protection 15:30 - Design system for consistent UI 15:50 - Sponsor message 18:00 - Building the AI Project Planner app 22:30 - Saving plans for future sessions 24:30 - Creating custom sub agents 28:30 - Playwright MCP server setup 30:30 - Auto testing the app with browser MCP Business & sponsorship enquiries: leon.vanzyl@gmail.com #opencode #agenticcoding

An LLM is an incredibly powerful brain that knows everything about humanity and nothing about you or the software you run. This video walks through the four buzzwords that turn that raw brain into a system you can trust: agent harness, loop engineering, LLM Ops, and eval. No jargon, step by step, technical or not. You can build anything. You can learn anything. 💪 Timeline 0:00 Intro: the 4 AI agent buzzwords 0:42 What an AI agent run actually is 1:52 The memory system: procedural, semantic, episodic 3:37 What "harness" really means (the horse) 5:07 Storing and updating memory (databases, skills, summarizer agent) 8:28 Retrieval: RAG, SQL vs semantic search 10:00 Tool calling and why agents loop 10:52 Loop engineering and end-loop guardrails 13:21 A Claude Code hooks example 14:18 Eval and LLM Ops: why you need them 15:55 Tracing every run (Langfuse, LangSmith) 16:43 Evaluation: LLM as a judge 17:20 Diagnosing what broke 18:16 The gate: ship the fix or fix the bug 18:57 Zoom out: the full system 19:53 Wrap up Launch Full-Stack App Template: https://github.com/ShenSeanChen/launch-mvp-stripe-nextjs-supabase My AI Agent Startup AutoManus.io is an AI sales lead manager for made-to-order product sales. We embed directly where conversations happen (WhatsApp, email, web chat) to capture inbound, automate follow‑ups, and kill CRM busywork; live with 4 pilots and 2 contracts in motion, pre‑seed backed by Character VC. Focused on US/UK/Germany, raising to expand engineering and scale deployments. ☕️ Buy me a coffee: https://buy.stripe.com/5kA176bA895ggog4gh - X: https://x.com/ShenSeanChen - Sean's Discord: https://discord.gg/tvECErKcFr - AutoManus Discord: https://discord.gg/5HhcNjCR - LinkedIn: https://linkedin.com/in/shen-sean-chen - Github Link: https://github.com/ShenSeanChen - Instagram: https://www.instagram.com/sean_ai_stories Sub Count when launched: 13,050 Top Github Repos - LaunchMVP NextJS,Supabase,Stripe (1015 stars): https://github.com/ShenSeanChen/launch-mvp-stripe-nextjs-supabase - ReactNextJSTips (114 stars): https://github.com/ShenSeanChen/yt-react-nextjs-tips - DeepResearchAgentBackend (70 stars): https://github.com/ShenSeanChen/yt-DeepResearch-Backend - SupabaseTipsKanban (64 stars): https://github.com/ShenSeanChen/yt-supabase-tips-kanban - RAGVectorDB (56 stars): https://github.com/ShenSeanChen/yt-rag - KimiK2Claude (37 stars): https://github.com/ShenSeanChen/yt-kimi-k2-claude - DeepResearchAgentFrontend (33 stars): https://github.com/ShenSeanChen/yt-DeepResearch-Frontend - AgenticRAG (13 stars): https://github.com/ShenSeanChen/yt-agentic-rag You Can Build Anything Playlist (Full-Stack Dev Without CS Degrees) - Core7: 95% React/NextJS Use Cases https://www.youtube.com/watch?v=vZzFlAjz4rA&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=16 - Core6: 95% Supabase Use Cases https://www.youtube.com/watch?v=Pnoq2tIif3s&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=15 - Core5: StripeNextJSSupabase https://www.youtube.com/watch?v=ad1BxZufer8&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=8 - Core4: AuthSupabaseGoogle https://www.youtube.com/watch?v=AZAU3ClAWWE&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=6 - Core3: FastapiGCPPro https://www.youtube.com/watch?v=u3sfqpzzo-U&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=5 - Core2: SupabaseHero https://www.youtube.com/watch?v=V0kmgkyPsVQ&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=4 - Core1: LaunchAnApp https://www.youtube.com/watch?v=iQ2izvGG4tQ&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=3 - Launch WhereToMeet w/ GoogleMaps, Calendar, Supabase, Cursor: https://www.youtube.com/watch?v=5xMjX1QxGUY&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR&index=2 - Launch EncourageMe w/ Cursor, Supabase, Vercel https://www.youtube.com/watch?v=hnvL4xa2DHA&list=PLE9hy4A7ZTmpGq7GHf5tgGFWh2277AeDR

Watch us build a complete real-time polling app using ONLY AI! Andrew Stewart (@srcrip) & Daniel Bergholz (@DanielBergholz) join me as we let Claude Code + Tidewave do ALL the coding. No manual code writing allowed! We built authentication, real-time voting, dynamic forms, live updates, and a polished UI - all through AI prompts. The result? A fully functional Phoenix LiveView app with 13 commits in just 2 hours! 00:00 - Introduction and sponsor message 04:00 - Setting up the live coding challenge and polling the audience 09:00 - Introducing Claude Code and the wild card project selection 15:25 - Installing and configuring Claude Code for the first time 23:30 - Setting up Tidewave MCP for Elixir integration 32:00 - Creating the Phoenix project and initial setup 40:00 - Building polls schema and embedded options 48:00 - Implementing Phoenix authentication with gen.auth 1:05:00 - Creating live views and dynamic forms 1:20:00 - Adding real-time voting functionality with PubSub 1:27:00 - Debugging with Tidewave's powerful runtime integration 1:35:00 - Building authorization and user permissions 1:45:00 - Adding vote changing functionality and UI polish 1:55:00 - Creating a stunning homepage with Claude's design skills 2:05:00 - Setting up GitHub workflow and creating pull request 2:15:00 - Wrap-up and reflections on AI-assisted development Tools Used: Claude Code (agentic terminal AI) Tidewave MCP (Elixir runtime integration) Phoenix LiveView + Tailwind CSS Judge Our Vibe Code: Open PR: https://github.com/ElixirMentor/pulse_vote/pull/1 Connect: Andrew: @srcrip Daniel: @DanielBergholz Support Elixir Mentor: https://elixirmentor.com/?utm_source=elixir-mentor #TideWave #ClaudeCode #ElixirLang #PhoenixFramework #AIcoding #VibeCoding #LiveView #ElixirMentor

🚀 Work with us - https://aifoundations.io/apply ✨ Join Ai Foundations - https://aifoundations.io This is the only Claude tutorial you’ll ever need. With this guide you can go from beginner to pro using Claude. I cover the basics like signing up, settings, memory, prompting and we work our way up to the intermediate features like Claude projects, artifacts, inline visualizations, data analytics and more. Then we finally dive into Claude desktop, where I teach you how to use Claude Cowork with computer use, scheduled tasks, and connectors for Notion, Gmail and Google Calendar. I even show you how to build Claude skills from scratch. After you have a beginner's understanding of Claude we dive into Claude Code. I show you how to deploy a real website in Netlify and even control Claude through a remote environment on Anthropics servers. We cover github, environment variables, pretty much everything you need, explained in beginner terms. I even cover exporting your ChatGPT memories and bringing them over to Claude. If you've been hearing about Claude AI, Claude Cowork, or Claude Code and wondering how it all fits together: this is the video you’ve been searching for. I walk through everything step-by-step leaving no stone unturned. Utilize the chapters and resources below to get the most out of this video. If it’s your first time watching, I recommend watching all the way through at least once, then coming back to spend extra time on specific chapters! Resources Mentioned: Join Ai Foundations - https://aifoundations.io DBS Framework Skill - https://aifoundations.io/resources/the-dbs-framework-skill Claude Code Mastery - https://aifoundations.io/products/claude-code-mastery Sign up for Claude - https://claude.ai Chapters: 0:00 - Claude Tutorial Intro 0:25 - How to Use This Guide 0:40 - Claude Pricing Comparison 1:22 - Sign Up for Claude 1:49 - Claude Onboarding 3:10 - Claude Chat Layout 3:50 - General Settings 4:53 - Notifications 6:00 - Claude Appearance 6:35 - Claude Voices 6:51 - Privacy and Billing 8:02 - Claude Memory 8:26 - Exporting Memories from ChatGPT 9:21 - Tool Settings 10:14 - Claude Artifact Settings 11:23 - Code Permissions 12:10 - Connector Settings 13:00 - Language Settings 13:15 - Claude Status and Support 13:50 - Gifting Claude 14:23 - Keyboard Shortcuts 14:40 - Claude Models 16:15 - Attachments and Connections 17:10 - Writing Styles in Claude 17:53 - Talking to Claude Voice 19:00 - Organizing Chats 19:40 - Incognito Chats 20:20 - GCAO Prompt Template 21:50 - Prompting Claude 24:50 - Ai Foundations Offers 25:38 - Claude Chat Search 26:34 - Claude Projects Guide 27:18 - Project Instructions 29:40 - Projects Guide Continued 30:56 - Claude Artifacts Library 32:37 - Creating Claude Artifacts 36:40 - Sharing Artifacts 37:15 - Mini App Artifacts 38:35 - Inline Visualizations 39:00 - Visualizing Data Analytics 41:15 - Claude Desktop 42:45 - Claude Cowork Setup 44:00 - Claude Quick Chat 45:00 - Cowork Settings Continued 47:00 - Organizing Files in Cowork 48:20 - Claude Computer Use 49:30 - Cowork Scheduled Tasks 51:22 - Cowork Projects 52:40 - Cowork Prospecting Project 55:00 - Claude in Chrome Extension 56:10 - Cowork Project Results 57:05 - Claude Dispatch Setup 57:50 - Cowork Connectors 58:20 - Claude Notion MCP Connector 59:30 - Using Connectors 1:00:18 - Claude Gmail Connector 1:02:40 - Claude Skills 1:04:18 - DBS Framework Skill 1:05:30 - Uploading Skills 1:06:00 - Creating a Basic Skill 1:07:36 - Testing the Recipe Skill 1:09:25 - Creating an Advanced Skill 1:10:20 - Google Calendar Connector 1:12:24 - Testing the Calendar Skill 1:13:10 - Claude Code Prerequisite 1:14:20 - Claude Code Mastery 1:15:23 - Two Ways to Use Claude Code 1:15:48 - Claude Code Environments 1:17:45 - Claude Code Models 1:18:10 - Claude Code Effort 1:18:34 - Claude Code Permissions 1:18:50 - Claude Code Enhancements 1:20:20 - Folder Setup 1:21:30 - Creating Claude MD 1:22:12 - Opening Files Quickly 1:22:31 - Building a Personal Website 1:25:15 - Planning with Claude Code 1:28:05 - Plan Mode in Claude Code 1:28:45 - Editing My Site 1:29:19 - The Complete Website 1:30:00 - Deploying with Netlify Connector 1:31:12 - Setting a Domain Name 1:31:45 - Adding Environment Variables 1:34:00 - Testing Email Connection 1:34:30 - Creating Remote Skills 1:36:18 - Github Repo Setup 1:37:00 - Initial Commit in Github 1:38:00 - Claude Code Cloud Environments 1:39:00 - Claude Code on my Phone 1:39:55 - Netlify + Github Deployment Setup 1:41:00 - Updating My Website 1:42:00 - Website Complete 1:42:20 - Claude Code Scheduled Tasks 1:44:24 - Scheduled Tasks with Skills 1:47:30 - Updating My Website from Anywhere 1:49:50 - Claude Code Terminal 1:50:30 - Claude Code Commands 1:51:10 - Claude Code Terminal Test 1:51:58 - Final Thoughts #claude #claudecowork #claudecode

Learn how to use the OpenAI's Assistants API to build powerful AI assistants. In this course, we'll explore how to leverage the Assistants API by OpenAI to create dynamic, intelligent web apps using Streamlit. What we'll cover: 1️⃣ Function Calling with the API: Learn to seamlessly integrate the Assistants API into your applications, enabling advanced AI functionalities right at your fingertips. 2️⃣ Knowledge Retrieval: Discover how to use the API to extract information, answer questions, and make your applications smarter and more responsive. 3️⃣ Code Interpreter Capabilities: Dive into the API's ability to interpret and generate code, a game-changer for automating tasks and enhancing your app's capabilities. 4️⃣ LLM Fundamentals: Gain a solid understanding of Large Language Models (LLMs) and how they form the backbone of OpenAI's Assistants API. This module demystifies the technology and provides a foundation for advanced application development. The course features a series of hands-on projects and real-world examples to apply what you've learned. By the end of our session, you'll have the skills and confidence to build your own intelligent web apps using Streamlit and OpenAI's Assistants API. ⭐️ Code ⭐️ Personal trainer: https://github.com/pdichone/vincibits-personal-trainer-assistant Personal trainer: https://github.com/pdichone/vincibits-personal-trainer-assistant Study-Buddy: https://github.com/pdichone/vincibits-study-buddy-knwoledge-retrieval ✏️ Created by @vincibits Twitter (X): @buildappswithme ❤️ Try interactive AI courses we love, right in your browser: https://scrimba.com/freeCodeCamp-AI (Made possible by a grant from our friends at Scrimba) ⭐️ Contents ⭐️ ⌨️ (0:00:00) Introduction ⌨️ (0:01:02) What’s This Course About - What Will You Learn? ⌨️ (0:01:33) Pre-requisites ⌨️ (0:02:44) Python and Dev tools Set up ⌨️ (0:04:22) VS Code Installation ⌨️ (0:05:31) OpenAI Account - Generate an API Key ⌨️ (0:07:53) What is the Assistants API & Benefits - Comparison Between the Chat Completion API and the Assistants API ⌨️ (0:18:16) Assistants API Building Blocks ⌨️ (0:24:04) Creating an Assistants API - Manually (Personal Trainer) ⌨️ (0:38:20) Creating an Assistants API - In Code (Personal Trainer) ⌨️ (1:04:15) Build a News Summarizer Introduction: Function Calling - A Streamlit Application ⌨️ (1:25:39) Create an AssistantsManager Class For our News Summarizer ⌨️ (2:10:46) Using the AssistantManager Class to Create an Assistant and run it as a Streamlit App ⌨️ (2:28:23) Knowledge Bases Retrieval Tools - How it Works & Introduction to Embeddings ⌨️ (2:35:25) Build a Study Buddy Streamlit Application ⌨️ (3:22:24) Run the Study Buddy Streamlit Application and Test ⌨️ (3:29:27) Wrap up and Final Considerations. 🎉 Thanks to our Champion and Sponsor supporters: 👾 davthecoder 👾 jedi-or-sith 👾 南宮千影 👾 Agustín Kussrow 👾 Nattira Maneerat 👾 Heather Wcislo 👾 Serhiy Kalinets 👾 Justin Hual 👾 Otis Morgan 👾 Oscar Rahnama -- Learn to code for free and get a developer job: https://www.freecodecamp.org Read hundreds of articles on programming: https://freecodecamp.org/news