CCAR-F · Claude Certified Architect – Foundations

Scenario × domain matrix

CCAR-F is the only scenario-based exam of the four. It draws 4 of the 6 published scenarios, so what you actually get asked is decided by that draw. This is the map: which domains each scenario reaches, and how much of the blueprint each one spans.

  • Guide v1.0 · effective 2026-07
  • 60 items · 120 min
  • Cut score 720 of 1000
  • 4 of 6 scenarios
  • Valid 12 months

Exam shape and bank readiness

  • Exam form

    60 items

    120 minutes — about 2.0 min per item.

  • Cut score

    720 of 1000

    Scaled 100–1000 from a standard-setting study. Nothing here predicts it.

  • Scenario draw

    4 of 6 drawn

    15 possible draws · ~15 items per scenario.

  • Draws covering all domains

    14 of 15 93%

    1 draw misses a domain entirely.

  • Item bank

    30 of 30 statements at floor

    Every statement has at least 3 items. 120 authored in total.

Coverage risk in the draw

4 of the 6 scenarios are presented, which is exactly 15 possible draws. 14 of those 15 exercise all 5 domains (93%).

The only exception is {S1, S2, S3, S4}, which omits D4 — D4 is primary only in S5 and S6 ( Claude Code for Continuous Integration and Structured Data Extraction) . Skip those and you leave 20% of the exam unprepared.

D5 Context Management & Reliability carries the lowest weight (15%) but the widest scenario spread — primary in 4 of the 6 scenarios, and reachable through 15 of the 15 draws. Lowest weight is not lowest exposure.

Widest single scenario: S4

Developer Productivity with Claude spans 65% of the blueprint through D1, D2, D3 — more than any other scenario in the bank. Build this lab first.

Open S4

Narrowest route in: D4

Prompt Engineering & Structured Output is worth 20% of the exam but is primary in only 2 scenarios ( S5, S6), so it is reachable through 14 of the 15 draws rather than all of them. That asymmetry is the trap: a low draw probability is not a low weight.

Open D4

The matrix — 6 scenarios × 5 domains

Ordered by weight spanned, so the top row is the highest-coverage lab. Domain columns carry the published weight and the item count a real form is expected to draw from that domain (60 items × weight).

CCAR-F scenario bank by domain. Rows are the 6 published scenarios ordered by the share of the blueprint they span; columns are the 5 domains with their published weights. A filled cell means the domain is primary for that scenario.
Scenario D1 Agentic Architecture & Orchestration 27% ~16.2 items D2 Tool Design & MCP Integration 18% ~10.8 items D3 Claude Code Configuration & Workflows 20% ~12.0 items D4 Prompt Engineering & Structured Output 20% ~12.0 items D5 Context Management & Reliability 15% ~9.0 items Weight spanned of 60 items
S4 Developer Productivity with Claude D1 is primary for S4, 9 items authored D2 is primary for S4, 6 items authored D3 is primary for S4, 5 items authored D4 is not primary for S4 D5 is not primary for S4 65%
S1 Customer Support Resolution Agent D1 is primary for S1, 9 items authored D2 is primary for S1, 5 items authored D3 is not primary for S1 D4 is not primary for S1 D5 is primary for S1, 6 items authored 60%
S3 Multi-Agent Research System D1 is primary for S3, 8 items authored D2 is primary for S3, 6 items authored D3 is not primary for S3 D4 is not primary for S3 D5 is primary for S3, 6 items authored 60%
S5 Claude Code for Continuous Integration D1 is not primary for S5 D2 is not primary for S5 D3 is primary for S5, 9 items authored D4 is primary for S5, 11 items authored D5 is not primary for S5 40%
S2 Code Generation with Claude Code D1 is not primary for S2 D2 is not primary for S2 D3 is primary for S2, 12 items authored D4 is not primary for S2 D5 is primary for S2, 8 items authored 35%
S6 Structured Data Extraction D1 is not primary for S6 D2 is not primary for S6 D3 is not primary for S6 D4 is primary for S6, 13 items authored D5 is primary for S6, 7 items authored 35%
Scenarios reaching this domain 3 /6 3 /6 3 /6 2 /6 4 /6 4 of 6 drawn
Possible draws reaching this domain 15 /15 15 /15 15 /15 14 /15 15 /15 15 draws
  • primary domain for that scenario
  • not primary
  • authored items for that scenario × domain pair
  • D1 Agentic Architecture & Orchestration
  • D2 Tool Design & MCP Integration
  • D3 Claude Code Configuration & Workflows
  • D4 Prompt Engineering & Structured Output
  • D5 Context Management & Reliability

Every number above is computed at build time from the published blueprint.yml for CCAR-F (guide v1.0). The possible-draw count is enumerated in code, not asserted, and the build fails if it disagrees with the published figure.