Certification Roadmap — to CCA-F

The single page that answers "what's between me and the certificate?" It threads the course's tracks through the five exam domains and ends at the proctored exam.

The destination

Claude Certified Architect — Foundations (CCA-F) — Anthropic's foundational technical certification. ~120 min, ~60 multiple-choice questions (1 correct + 3 distractors), proctored on Skilljar, drawing on five weighted domains and a pool of six scenarios. Full details: cca-f-exam-overview.

The route (course → exam)

flowchart TD
    A[On-ramp: Track 0A / 0B] --> B[Track 1 Foundations]
    B --> C[Track 2 Prompting]
    C --> D[Track 3 API]
    D --> E[Track 4 Claude Code]
    E --> F[Track 5 MCP & Tools]
    F --> G[Track 6 Agents/Skills/Subagents]
    G --> H[Track 7 Deployment]
    H --> I[Track 8 Certification: capstone + drills]
    I --> J{Readiness ≥ 80%?\nno domain < 70%}
    J -- yes --> K[Schedule & sit CCA-F]
    J -- no --> I

Domains → where you learn them

CCA-F domain	Weight	Covered by
Agentic Architecture & Orchestration	27%	Track 6 + introduction-to-subagents, p14-subagent-research, p15-multi-agent-orchestrator, agentic-patterns
Tool Design & MCP Integration	18%	Track 5 + introduction-to-mcp, mcp-advanced-topics, p12-build-an-mcp-server, tool-design-principles
Claude Code Config & Workflows	20%	Track 4 + claude-code-101, claude-code-in-action, p10-claude-code-onboarding, p11-custom-command-and-hook
Prompt Engineering & Structured Output	20%	Track 2 + prompt-engineering-interactive-tutorial, real-world-prompting, prompt-evaluations, p07-structured-extraction
Context Management & Reliability	15%	Track 3 + building-with-the-claude-api (caching, batch, error handling), p08-prompt-caching-and-rag, cca-domain-5-context-reliability

Milestones

Fluent user — Tracks 0–1 done; you use Claude confidently. (Optional: claim Anthropic's Claude 101 / AI Fluency certificates.)
Builder — Tracks 2–3 done; you can prompt reliably and build on the API.
Agentic engineer — Tracks 4–6 done; Claude Code, MCP, skills, subagents.
Deployer — Track 7 done; Claude on a cloud with governance.
Capstone — p16-capstone-support-agent complete end-to-end.
Drill & sit — Track 8: practice tests until readiness ≥ 80% with no domain < 70%, then schedule the exam with the cca-f-exam-coach.

Readiness signals (don't book early)

Overall readiness ≥ 80% in progress / learner-db.json.
No single domain below 70%.
You can walk all six scenarios from memory: architecture, trade-offs, three pitfalls each.
You've scored ≥ 80% on at least two full mock exams (see cca-f-exam-coach MOCK EXAM mode).

Day-of

Sleep, water, calm. Two hours; you can flag and review.
Read each question twice. On single-trade-off questions, the "obvious" answer is often the trap.
Use cheatsheet for the final-week brain-dump.