Claude Code Configuration & Workflows Quiz - Certifications Claude Architect Foundations

Free practice quiz: Claude Code Configuration & Workflows (Certifications Claude Architect Foundations). 78 multiple-choice questions with answers and explanations. Play online, no sign-up needed.

Sample questions

1. What is the most likely reason the new team members do not get these instructions?

   1. Because they are in a user-level CLAUDE.md file not shared via version control.
   2. Because the new members need to run /init to generate their own CLAUDE.md first.
   3. Because they were not referenced with @import from the project CLAUDE.md.
   4. Because project instructions must be placed in CLAUDE.local.md to be shared.
   Show answer

   Answer: A. Because they are in a user-level CLAUDE.md file not shared via version control.

   User-level memory applies only to that user's sessions and is not shared with teammates. Team-shared instructions must be placed in a project-level CLAUDE.md that is committed to version control.

2. What configuration change most directly prevents this skill's output from polluting the main conversation while still preserving its results?

   1. Add allowed-tools: Read to keep the skill from editing files.
   2. Run /compact immediately after each skill invocation.
   3. Move the skill's instructions into CLAUDE.md.
   4. Add context: fork to the skill frontmatter.
   Show answer

   Answer: D. Add context: fork to the skill frontmatter.

   context: fork runs the skill in an isolated sub-agent context and returns only a summarized result to the main conversation, so verbose output does not pollute the session. allowed-tools restricts which tools the skill may use but has no effect on output volume. Running /compact after the fact or moving the skill into CLAUDE.md does not isolate the skill's output.

3. What is the best way to apply the single API error convention to all these TypeScript service source files while avoiding duplicate instructions?

   1. Create .claude/rules/api-errors.md with paths: ["services/**/src/**/*.ts"]
   2. Add the API error convention to the repository root CLAUDE.md
   3. Create an api-errors skill and rely on Claude to load it when editing services
   4. Create one CLAUDE.md at services/CLAUDE.md with the API error convention
   Show answer

   Answer: A. Create .claude/rules/api-errors.md with paths: ["services/**/src/**/*.ts"]

   A path-scoped .claude/rules file with a TypeScript glob applies the convention only to matching service TypeScript files, avoiding duplication and irrelevant loading. A services-level CLAUDE.md covers all files under services, root CLAUDE.md loads unconditionally for every session, and a skill relies on model relevance rather than deterministic path matching.

4. What is the most effective first step with Claude Code?

   1. Enter plan mode and ask Claude to explore billing boundaries and dependencies, then compare both approaches before editing.
   2. Add a CLAUDE.md instruction that billing is now a separate service, then ask Claude to extract directly against the monolith.
   3. Use an Explore subagent to list all billing-related files, then ask Claude to extract using that report as the only guide.
   4. Ask Claude to implement the lower-risk strangler-fig migration approach directly and adjust it as integration issues appear.
   Show answer

   Answer: A. Enter plan mode and ask Claude to explore billing boundaries and dependencies, then compare both approaches before editing.

   Plan mode is designed for complex tasks involving large-scale changes, multiple valid approaches, architectural decisions, and multi-file modifications. Exploring and planning before coding prevents costly rework from choosing the wrong extraction pattern. Direct execution or a summary-only subagent cannot safely handle the architectural tradeoffs.

5. What is the most effective way to surface design considerations before implementation?

   1. Instruct Claude to use a 24-hour TTL on all search results, since that is the standard default for web search caches.
   2. Ask Claude to interview you with questions about cache invalidation, failure modes, and data freshness before writing any code.
   3. Ask Claude to generate a full system design document with recommended cache eviction policies, then review and correct anything that looks wrong.
   4. Implement the happy path first, then ask Claude to add a cache once search results start returning stale data.
   Show answer

   Answer: B. Ask Claude to interview you with questions about cache invalidation, failure modes, and data freshness before writing any code.

   The interview pattern surfaces design considerations the developer may not have anticipated. It is especially useful in unfamiliar domains such as agent caches where failure modes are not obvious.

More quizzes

• Prompt Engineering & Structured Output
• Tool Design & MCP Integration
• Agentic Architecture & Orchestration

Browse all Claude Architect Foundations quizzes