CLAUDE MODELS
Claude Models Guide: Choosing Opus 5 or Sonnet 5
A practical guide to choosing between Claude Opus 5 and Claude Sonnet 5 for reasoning, coding, agents, throughput, and cost-sensitive production work.
Open guide ↗XTIANZ CLAUDE DESK
CLAUDE BY ANTHROPIC
Build with Claude—without skipping the engineering discipline.
A practical guide to model choice, Claude Code, tool-using agents, MCP, enterprise governance, safety, and how Claude is changing work.
CLAUDE TOPIC MAP
Six areas that matter in production.
Use the cards as a learning path—from choosing a model to governing real enterprise deployment.
CLAUDE GUIDES
Practical, source-linked reading.
Each guide separates official product facts from XTIANZ engineering interpretation.
CLAUDE CODE
Claude Code Engineering Guide: Workflows, Hooks, MCP and the Agent SDK
How to use Claude Code as an engineering system: repository context, plans, tests, hooks, MCP connections, security controls, and the Agent SDK.
Open guide ↗CLAUDE AGENTS
Building Claude Agents with Tools, MCP, Web Search and Computer Use
A systems guide to building Claude agents with tool use, MCP, web search, computer use, prompt caching, evaluations, and operational controls.
Open guide ↗CLAUDE ENTERPRISE
Claude Enterprise Governance: Identity, Compliance, Security and Rollout
A practical governance framework for deploying Claude and Claude Code across an enterprise with identity, compliance, logging, retention, and operating controls.
Open guide ↗CLAUDE SAFETY
Claude’s Constitution and What It Means for AI Safety Engineering
What Claude’s constitution is, what it can and cannot guarantee, and how engineering teams should translate model-level safety into system-level controls.
Open guide ↗CLAUDE + WORK
Claude and the Future of Work: Reading the Anthropic Economic Index Carefully
How to interpret the Anthropic Economic Index and its Claude connector without turning task-level data into unsupported predictions about entire jobs.
Open guide ↗XTIANZ CLAUDE LENS
Model quality is only one layer.
A reliable Claude system also needs explicit tools, identity, permissions, evaluations, tracing, recovery, cost controls, and human accountability.
SELECTCONSTRAINEVALUATEOPERATE