CCA-Foundations colleague pass intel (score 902)
A phData colleague passed Claude Certified Architect – Foundations and posted a guide internally; founder forwarded it. This is the first concrete registration path AND the first insider difficulty map.
Registration (the gate-clearer)
- Since July 1: exam sits via Pearson VUE.
- Path: create account at anthropic-partners.skilljar.com using phData email → follow registration instructions at
anthropic-partners.skilljar.com/claude-certified-architect-foundations-certification/486716. - This likely dissolves the "partner-portal access" blocker open since 2026-06-28 — self-serve, no manager dependency. Founder action: create the account (~15 min).
- Logistics tip from the colleague: run the "test your system" check from the registration email well in advance — multiple people hit issues at the system check.
Difficulty map (60 questions / 120 minutes)
| Area | Difficulty (colleague's read) | Founder fit |
|---|---|---|
| Session management (resume/continue/fork) | Low-Med | daily bread — skip |
| Context & subagents (offloading, context degradation) | Med | strong (runs this architecture) |
| Task decomposition (chaining vs plan mode vs TDD vs direct) | Med-High, BIGGEST CHUNK | drill — volume alone makes it priority #3 |
| MCP fundamentals (tools vs prompts vs resources, discovery) | Med | solid |
| CLAUDE.md / Skills / Slash commands (project vs personal, loaded vs invoked) | Med | daily bread |
| Hooks & deterministic enforcement (code vs prompt) | Med-High | strong (his harness thesis) |
| Tool/error design for agents (structured errors, instructive messages) | HIGH | drill priority #2 |
| Escalation & human-in-the-loop | HIGH, very scenario-heavy | drill priority #1 — this IS the D1 act-vs-surface weak axis from the [[2026-06-27-cca-practice-exam-2-miss-analysis]] |
| Agentic loop mechanics (stop_reason, tool-result flow) | Low-Med | covered by [[2026-06-14-cca-harness-mechanics-cheatsheet]] |
The targeting insight: the exam's two HIGH-difficulty areas overlap his known weak axes almost exactly. Escalation/HITL scenario judgment = the D1 act-vs-surface-to-user misses; "which is BEST" test-craft applies most in the scenario-heavy sections. His Low-Med areas are his day job — zero drill time there.
Prep stack the colleague used
- phData enterprise Udemy: "Certified Claude Architect Masterclass 2026" (phdata.udemy.com)
- Official Anthropic documentation
- Volume of practice tests + choice-elimination technique ("in most questions you can eliminate the clearly wrong choices")
- Third-party mock: claudecertificationguide.com/mock-exam — unvetted third-party site; use for question-pattern feel only, don't treat its answer keys as authoritative
Related
- [[cert-progress]] — gate row updated 2026-07-13
- [[2026-07-02-cca-exam-guide-update-analysis]] — official scope; the out-of-scope list still governs what NOT to study
- [[2026-06-27-cca-practice-exam-2-miss-analysis]] — the two weak axes
- [[2026-06-14-cca-harness-mechanics-cheatsheet]]