/Stream
All 1133 resources sorted by date, newest first
Build Agents, Not Pipelines
Pipelines use fixed logic for predictability at scale; agents manage their own control flow for complex reasoning — a practical guide to knowing which to reach for
Chrome DevTools MCP Server Stable in Chrome 149
The MCP server and CLI for Chrome DevTools letting AI coding agents drive debugging and perf audits is now officially stable
How We Contain Claude Across Products
Anthropic's multi-layered containment approach — sandboxes, VMs, strict egress controls, and model training — limits agent blast radius more reliably than human supervision alone
Boxes
Cloud workbench for running multiple AI coding agents in parallel within isolated, persistent virtual machines — monitor and manage from any device
Code Wiki
Gemini-generated repo documentation that stays up-to-date — generates natural language summaries and architecture diagrams for any repository
Mnemo
Local-first AI memory layer that stores persistent knowledge in a SQLite knowledge graph and injects relevant context into LLM sessions automatically
Open Code Review
AI-powered code review CLI that reads git diffs, sends files to a configurable LLM via an agent with tool-use capabilities, and generates structured review comments with line-level precision
Paseo
Self-hosted, privacy-focused platform for orchestrating and running multiple AI coding agents across mobile, desktop, and CLI environments
Sandcastle
Provider-agnostic TypeScript library for running AI coding agents in isolated sandboxes — Docker, Podman, or Vercel
Agentic Engineering Patterns
A growing guide collecting coding practices and patterns for getting the best results from coding agents
I Didn't Become a Developer to Review AI Slop
AI made it cheap to open pull requests but not trustworthy, burying developers in mystery diffs that look clean on the surface but hide bugs and debt
Introducing Dynamic Workflows in Claude Code
Claude can now write orchestration scripts that spin up hundreds of parallel subagents for complex end-to-end tasks, in research preview for Max and Enterprise plans
Microsoft's Quiet Claude Code Retreat and the Real Cost of Enterprise AI
Microsoft cancels most Claude Code licenses in its Experiences and Devices group telling engineers to migrate to Copilot CLI by June 30
FLUE
A TypeScript agent harness framework from the Astro team, akin to Claude Code with tasks, sandboxing and agent skills but entirely headless and programmable
Resend Claude Code Plugin
Official Claude Code plugin with skills for creating React Emails, sending and receiving with Resend, agent inbox setup, and built-in Resend MCP server
Retune
Visual design tool that lets you select and modify website elements directly in the browser while an AI agent writes the corresponding code, integrating with Claude Code and Cursor
Skybridge
Full-stack React framework for type-safe MCP applications and servers, with hot module reloading and a local emulator for AI environments like Claude and ChatGPT
Webwright
SWE-style browser agent framework that gives agents a terminal to launch browser sessions and complete long-horizon web tasks end-to-end
AI's Impact on Software Engineers in 2026: Part 2
Survey of 900+ engineers — AI amplifies existing culture, codebase quality is dropping while management focuses on output, and junior devs are struggling most
The Last Six Months in LLMs
Coding agents became reliable daily tools and local open-weight models reached cutting-edge performance — Simon Willison's 6-month review