Domain 7 — Claude Code (3.1%)
3.1% of the exam, roughly 2 questions. Deliberately light — this is the Architect exam's territory. Learn the shape, not the depth.
7.1 What lives where
| File | Holds |
|---|---|
CLAUDE.md / .claude/CLAUDE.md | Project instructions: build commands, conventions, architecture. Context, not enforcement. |
CLAUDE.local.md | Your personal project notes; gitignored. |
.claude/rules/*.md | Topic files; with paths: frontmatter they load only when Claude touches matching files. |
.claude/settings.json | Shared permissions, hooks, environment. Enforcement. |
.claude/skills/<name>/SKILL.md | A packaged procedure, loaded on demand. |
.claude/commands/*.md | Custom slash commands. |
.mcp.json | Project-scoped MCP servers, committed for the team. |
The one conceptual point that carries marks: CLAUDE.md is context, not
configuration. It shapes behaviour and guarantees nothing. A rule that must always hold belongs
in permissions.deny or a hook.
7.2 Interaction modes
Permission modes set the approval baseline: default prompts on first use of each tool;
plan explores read-only and produces a reviewable plan before any edit; acceptEdits
auto-accepts file edits; bypassPermissions skips prompts and belongs only in isolated
environments. Plan mode before a large refactor is the habit the exam rewards.
7.3 Running it programmatically
claude -p "Review this diff for security issues" \
--output-format json \
--allowedTools "Read,Grep" \
--bare
-p runs non-interactively; --output-format gives text,
json (result plus session ID and usage) or stream-json; --bare skips
auto-discovery of hooks, skills, plugins, MCP servers and CLAUDE.md, which is what makes a CI
run reproducible across machines. Exit code 0 on success, non-zero on failure.
← Previous: Domain 6 — Security and Safety | Next: Domain 8 — Eval, Testing and Debugging →
Explore Other Certification Study Guides
Switch to another exam guide to prepare for your next Claude certification:
Claude Certified Associate — Foundations
Entry-level certification for non-coders: prompting, output evaluation, model selection, workflow integration, configuration, governance & responsible use.
Claude Certified Architect — Foundations
Core architectural principles: agentic loops & orchestration, tool design & MCP, Claude Code workflows, prompt engineering, and context reliability.
Claude Certified Architect — Professional
Advanced enterprise architecture: solution design, system integration, evaluation frameworks, governance, safety & developer enablement.