Building Claude Agents with Tools, MCP, Web Search and Computer Use | XTIANZ

Skip to content

XTIANZAI SYSTEMS

Intelligence

SignalsCurrent evidence briefs MarketsAI economics and capital DMV PulseLocal infrastructure and news

Build

ArchitecturePatterns and system layers AI + MCPProtocols, agents and security ClaudeModels, code and agents AI LabInteractive engineering tools AI RiskReadiness and controls

Start here

Home Current signals

Build AI systems

Architecture AI + MCP Claude AI Engineering Lab AI Risk

Track the ecosystem

MANGOS companies Markets DMV Pulse

XTIANZ

About Editorial policy Corrections

CLAUDE AGENTS

AGT

On this page

The Claude agent loop
Tool use
MCP as the connection layer
Web search and computer use
Prompt caching
Production controls
What to measure

More Claude guides

Claude Models Guide: Choosing Opus 5 or Sonnet 5
Claude Code Engineering Guide: Workflows, Hooks, MCP and the Agent SDK
Claude Enterprise Governance: Identity, Compliance, Security and Rollout
Claude’s Constitution and What It Means for AI Safety Engineering
Claude and the Future of Work: Reading the Anthropic Economic Index Carefully

Claude at XTIANZ

Models, Claude Code, agents, enterprise controls, safety, and the future of work.

Open Claude hub

Home Signals Lab

Search XTIANZ

Jump to a section or search the XTIANZ library

AI systems intelligence for people building, evaluating, and tracking the infrastructure behind artificial intelligence.

Reviewed July 29, 2026 · Educational content · Primary-source links for changing claims.

IntelligenceSignals Markets DMV Pulse

EngineeringArchitecture AI + MCP Claude AI Lab AI Risk

CompanyMANGOS About Editorial policy

LegalCorrections Privacy Disclaimer Sitemap

© 2026 XTIANZ