Claude Developer Certification Domain 8: Tool Use, Tool Choice, MCP Servers & Agent Customisation

Domain 8 of the Claude Certified Developer (Foundations) exam prep series: Tools & MCPs (10.6% of the exam). A narrated, annotated walkthrough of how to define tools Claude chooses well, run the tool-use loop correctly, handle errors, and connect systems through the Model Context Protocol, updated for current Claude models. What you'll learn: • Anatomy of a tool definition and writing great tool descriptions • The client tool-use loop: tool_use, tool_result and matching IDs • tool_choice options, parallel tool calls, and why forced tool use now fails on the newest models • Reporting tool errors with is_error • Client tools vs. server tools (web search, web fetch, code execution), tool search and programmatic tool calling • Model Context Protocol: hosts, clients, servers, tools, resources and prompts • MCP transports (stdio and Streamable HTTP), building a server, and connecting it to Claude • Agentic customisation: tool guidelines, chaining, CLAUDE.md, skills, subagents and hooks Chapters: 00:00 Domain 8: Tools & MCPs 00:42 Anatomy of a Tool Definition 01:31 The Client Tool-Use Loop 02:26 Tool Choice & Parallel Calls 03:20 Tool Results & Errors 04:08 Client Tools vs. Server Tools 05:14 The Model Context Protocol 06:10 MCP Transports & Servers 07:13 Agentic Customisation 08:16 Domain 8: Knowledge Check This is an independent study resource and is not affiliated with or endorsed by Anthropic. Always check the official Claude documentation (platform.claude.com/docs) for the latest details. #Claude #MCP #ToolUse #Anthropic #AIAgents