Examworthyexamworthy.comClaude Certified Architect - Foundations cheat sheet
Anthropic
Free to share. Examworthy is not affiliated with or endorsed by Anthropic; CCAR-F and related marks belong to their respective owners.
At a glance
Format: Proctored multiple-choice and multiple-response, scenario-based (4 scenarios drawn from a bank of 6)
Domain weight map
Heaviest first - spend your time hereHow this exam thinks
CCAR-F tests whether you can tell a mechanism that shifts a model's behaviour from one that constrains it, and pick the right one for the guarantee the scenario asks for.
Spot the trap
Tempting wrong answers, and why they failTempting but wrong
Tool call ordering between two MCP tools is enforced by the server that publishes them.
Why it fails
MCP servers do hold connection state, but the protocol exposes independent tools with no cross-tool ordering semantics. There is no server-side sequencing rule that could lapse and explain out-of-order calls.
Agentic Architecture & Orchestration
Tempting but wrong
A PostToolUse hook that logs every migration command with its target connection string, reviewed weekly, prevents the outage.
Why it fails
The logging is genuinely useful for accountability, but it fires after the command has already run, so it improves detection of the outage rather than preventing it.
Claude Code Configuration & Workflows
Tempting but wrong
A schema requiring file and line fields guarantees that every finding came from the reviewed pull request's diff rather than the wider repository.
Why it fails
The schema only requires a string and an integer in those fields, not that the position falls inside the changed lines. Confirming a finding sits within the diff needs a separate check the pipeline is not running.
Prompt Engineering & Structured Output
Tempting but wrong
A PostToolUse hook that alerts an on-call engineer whenever process_refund is called meets an incapability requirement.
Why it fails
A hook is the right family of mechanism and gives a deterministic signal, but PostToolUse runs after the call has executed, so the refund has already been paid by the time the alert fires.
Tool Design & MCP Integration
Tempting but wrong
Tool results returned after an injected fact overwrite it in the conversation state.
Why it fails
Tool results dominate recent context by volume, but messages accumulate rather than overwrite one another - the block stays in the request. Volume dilutes attention; it does not delete content.
Context Management & Reliability
Tempting but wrong
Schema validation guarantees that the model resolved the correct order the customer was referring to.
Why it fails
Validation says nothing about which order was meant; it only checks structure. The server refunded exactly the identifier it was handed, so the fault is not in how the server resolved it.
Agentic Architecture & Orchestration
Tempting but wrong
Ask the agent in a fresh session to list the rules it is currently working under, and compare that list against the repository's topic files.
Why it fails
This appears to read the loaded configuration directly, but it is a self-report generated from context rather than an inspection of the harness, so it can omit or invent entries with no signal that it has.
Claude Code Configuration & Workflows
Tempting but wrong
A finding about code outside a pull request's diff is a false positive because it cannot be verified against the pull request under review.
Why it fails
The finding is verifiable against the file it names; the real problem is that acting on it falls outside the pull request author's remit. That is a scope issue, not an accuracy issue, even though defining correctness relative to the gate is tempting.
Prompt Engineering & Structured Output
Key terms
Exam-day rules
- Find the guarantee in the stem before you read the options. Words such as must, every time, before any, or a quoted residual failure rate mean a deterministic mechanism is required, and every probabilistic option is a distractor no matter how well written.
- Read the last sentence of the stem first. The scenarios are long and the requirement is usually in the final line, so knowing it before you read the setup turns a wall of context into a search.
- Check how many responses the item asks for. Multiple-response items state the number, and selecting the wrong count is a self-inflicted loss on questions you actually knew.
- When two options both look right, prefer the one that keeps information recoverable: structured error metadata over a generic status, attributed conflicting figures over a merged number, an explicit null over a silently missing field.
- Ask what the agent can actually see at that point. A large share of wrong options assume a subagent inherited context, that a compacted session still holds a figure, or that a session can objectively review its own output.
Revision schedule
- Day 1Read the exam guide and inventory the six scenarios
- Week 1Build a working agentic loop from scratch
- Week 2Design and break a set of tools
- Week 2Configure Claude Code properly on a real repository
- Week 3Drill structured output and its limits