Match a job Paths Subjects Questions Quizzes Pricing
Overview Read Practice

Practice — MCP Servers and Subagent Orchestration (15 questions)

Pro content

Sign up free, then start a 14-day Pro trial — no card needed.

Advanced Open Pro

Scoping Credentials for a New MCP Server

Unlock this question →
Advanced Open Pro

Deciding When a Subagent Is Worth It for a Single Task

Unlock this question →
Advanced Open Pro

Parallel Fan-Out: Wall Clock vs. Token Cost

Unlock this question →
Advanced Open Pro

Diagnosing a Clobbered Refactor

Unlock this question →
Advanced Open Pro

Designing a Multi-Agent Review Pipeline for a Large PR

Unlock this question →
Advanced Open Pro

An Agent That Cleaned Up a Little Too Well

Unlock this question →
Advanced Open Pro

Getting an OAuth-Gated MCP Server Working in a Headless CI Job

Unlock this question →
Advanced Open Pro

Naming, Resuming, and the Right Execution Mode for a Multi-Stage Subagent Task

Unlock this question →
Advanced Open Pro

Which MCP Primitive Puts the Model in Control

Unlock this question →
Advanced Open Pro

What `managed-mcp.json` Actually Locks Down

Unlock this question →
Advanced Open Pro

Which Built-In Subagent Type You Can Actually Resume

Unlock this question →
Advanced Open Pro

Do Subagents Background Themselves in a Headless `-p` Run

Unlock this question →
Intermediate Open Free

Why a Subagent's Investigation Doesn't Bloat Your Session Permalink →

You ask Claude Code to "use a subagent to investigate how our authentication system handles token refresh" instead of investigating directly yourself.

Why does this keep your main session's context small, even if the investigation reads dozens of files?

Share this question

Intermediate Open Free

The Subagent Claude Code Delegates Searches To Permalink →

While handling a request, Claude Code decides a broad codebase search is needed and dispatches it to a specific built-in subagent rather than running the search itself in the main context.

Which subagent is this?

Share this question

Intermediate Open Free

Why a CLI Tool Beats an Equivalent MCP Server Permalink →

Your project has both an MCP server and a CLI (gh) for talking to GitHub, and both can do the task you need.

Per this subject's context-management guidance, which should you prefer, and why?

Share this question

We use cookies for product analytics to improve OmniAtlas. See our Privacy Policy.