Blog Posts
Official Anthropic engineering, product, and enterprise AI updates. This collection now includes 29 articles covering Claude Code, Agent architecture, Context Engineering, MCP, production safety, enterprise adoption, and AI-native startups.
New in This Update
- Introducing Claude Opus 4.8 - Opus upgrade with stronger coding, agentic task performance, dynamic workflows, and effort control.
- How we contain Claude across products - Anthropic's engineering approach to limiting Agent blast radius across claude.ai, Claude Code, and Cowork.
- KPMG integrates Claude across its global business - A large-scale enterprise AI rollout across 276,000+ professionals.
- Anthropic acquires Stainless - Why SDK and API quality matter for the Agent ecosystem.
- PwC deploys Claude for enterprise functions - Claude Code and Cowork moving into enterprise technology build, deals, and business functions.
Highlights
Agent Architecture & Design
- Building Effective Agents - Anthropic's classic guide on agent design.
- Multi-Agent Research System - Orchestrating multi-agent collaboration.
- Scaling Managed Agents - From prototype to production.
Claude Code in Practice
- Claude Code Best Practices - Patterns for large codebases.
- Claude Code Auto Mode - Two-layer defense security.
- Claude Code Sandboxing - Code execution security.
Context & Tools
- Effective Context Engineering - Core context management in the agent era.
- Equipping Agents with Skills - Three-level progressive disclosure.
- Writing Effective Tools for Agents - Tool design principles.
Evaluation & Analysis
- Demystifying Evals for AI Agents - Eval framework design.
- BrowseComp: Eval Awareness - Browsing capability evaluation.
Startup Guide
- The Founder's Playbook - AI-native startup lifecycle.