Tool Design & MCP Integration
18%Designing tool interfaces, structured error responses, MCP server integration, and leveraging Claude Code's built-in tools.
Designing clear, composable tool interfaces that Claude can use effectively. Tool descriptions are the primary mechanism for tool selection -- minimal or ambiguous descriptions lead to unreliable routing, while well-crafted descriptions with input formats, examples, and boundaries dramatically improve accuracy.
Returning structured errors from MCP tools so Claude can recover gracefully. The isError flag, error categories, retryable vs non-retryable distinctions, and how subagents should implement local recovery for transient failures while propagating only unresolvable errors.
Controlling which tools Claude uses and when. Too many tools (18+) degrade selection reliability. Scoping tools to agent roles with 4-5 per agent, using tool_choice modes (auto, any, forced) to control behavior, and preventing agents from misusing tools outside their specialization.
Building and connecting MCP servers for Claude Code and the Anthropic ecosystem. Project-level vs user-level configuration, environment variable expansion for secrets, tool discovery at connection time, MCP resources for content catalogs, and choosing community servers vs custom implementations.
Understanding Claude Code's six built-in tools and when to use each one. Grep for content search, Glob for file path matching, Read/Write for full file operations, Edit for targeted modifications, and Bash for shell commands. Includes the incremental codebase exploration pattern: Grep entry points, Read imports, trace flows.