A phased, resource-by-resource plan built around the actual exam blueprint — plus a trainer for all 30 task statements with scenario questions in the real exam style.
60Questions
120 minTime limit
720/1000Pass mark
4 of 6Scenarios
$125Fee (USD)
Roadmap readiness
0%
0 of 0 steps done — tick tasks below as you go (saved in this browser).
Straight answer on "guaranteed": nothing — no course, mock, or roadmap — can guarantee a pass on a proctored, scenario-based exam. What follows is the highest-probability path: learn by building, drill the weighted domains, then sit mocks under real conditions. It is criterion-referenced (you're measured against a fixed standard, not a curve), so if you can genuinely make the calls in the trainer below, you pass.
Your head start: you run Claude Code daily with custom skills, subagents, MCP servers and CLAUDE.md hierarchies. That's ~65% of the exam weight (Agentic Architecture + Claude Code Config + Tool Design/MCP) already in your muscle memory. Focus the trainer on your likely-weakest domains — Domain 4 (structured output / JSON schema / batch) and Domain 5 (context, escalation, provenance).
01
Know the target
Five domains, weighted (from the official blueprint). Study time should follow the weights — Agentic Architecture alone is more than a quarter of the exam. Every question is a real-world scenario asking for the best decision, never a definition.
Six phases, in order. The sequence matters: build real intuition first, drill scenarios second, sit the exam last. Trying to memorize is the #1 reason people fail this one.
01
Orient & baseline
~2 hrs · Day 1
Find out where you stand before spending a study hour, and lock down access.
Anthropic asks CCAR-F candidates to complete four Partner Academy courses. They're split into modules and lessons — with per-lesson tracking and a section quiz to clear each one — on the dedicated learning page.
The scenarios are lifted straight from how Anthropic thinks about agents. Read the source — the "best decision" answers live here.
04
Do the guide's 4 exercises + drill the trainer
~6–8 hrs · Week 2
Section 8 of the guide lists 4 hands-on exercises that reinforce every domain. Building them turns knowledge into the intuition the scenarios test.
05
Mock exams under real conditions
~6 hrs · Week 3
Timed, closed-book, no AI — like the real thing. Review every wrong answer and re-drill its domain. Mocks find gaps; they aren't for a vanity score.
06
Final 48 hours & exam day
~2 hrs · +rest
Light touch. Consolidate weak spots, confirm logistics, then stop cramming.
03
Task statement trainer
Work the roadmap above first — these scenario questions assume the foundation it builds, so a newcomer shouldn't jump straight to Task 2. Then go through all 30 task statements from Section 6 of the official guide: expand each for a plain-English "what you need to be able to do," test yourself with an exam-style question (reveal the answer + reasoning), and tick "I can make this call" when it's solid. The meter tracks coverage per domain. Don't memorize these — self-assess against them.
Domain mastery
0 / 30
04
Full resource library
Official = made by Anthropic; Community = free but unofficial — good for practice, not authoritative on exam specifics.