Certification
    June 9, 2026

    CCA · Page 7 — The 6 Exam Scenarios

    All six CCA Foundations exam scenarios mapped to the domains they test, the decisions probed, and the anti-patterns baked into the wrong answers.

    Share

    The 6 Exam Scenarios

    CCA Foundations course · Page 7 of 8 · ← Back to all courses · 4 of these 6 scenarios appear on each exam. For each, learn the key decisions and the anti-patterns (every wrong answer is built from those).


    1. Customer Support Resolution Agent

    Domains: D1, D2, D5 Key decisions: when to escalate vs attempt resolution · structured MCP errors with isError/errorCategory · PreToolUse hook for a refund threshold · context preservation across long conversations · case-facts block for exact amounts. Anti-patterns (wrong answers): silent suppression of timeouts · sentiment-based escalation · generic error responses.

    2. Code Generation with Claude Code

    Domains: D3, D5 Key decisions: CLAUDE.md hierarchy and where standards live · plan mode for complex tasks · skills with context: fork for verbose workflows · independent review instance. Anti-patterns: self-review by the same session · team standards in personal config · /compact losing exact values.

    3. Multi-Agent Research System

    Domains: D1, D2, D5 Key decisions: narrow-decomposition bug → fix the coordinator, not the subagents · parallel Task calls in one response · explicit context to subagents · provenance with source + date · gap annotation. Anti-patterns: direct agent-to-agent communication · assuming context inheritance · picking/averaging conflicting stats.

    4. Developer Productivity with Claude

    Domains: D2, D3, D1 Key decisions: Grep (contents) vs Glob (paths) · MCP vs built-in tool descriptions · CLAUDE.md for team standards · session management during long tasks. Anti-patterns: few-shot examples to fix misrouting (fix descriptions instead) · team standards in ~/.claude · resume on stale context.

    5. Claude Code for CI/CD

    Domains: D3, D4 Key decisions: -p flag required to prevent hanging · multi-pass architecture for 14+ files · independent review instance with zero prior context · prior findings to prevent duplicate comments. Anti-patterns: missing -p flag · single pass on large PRs · self-review same-session bias.

    6. Structured Data Extraction

    Domains: D4, D5 Key decisions: nullable fields + remove from required for absent data · self-correction pattern with conflict_detected · retry for format errors only · Batch API for high-volume async · provenance for claims. Anti-patterns: required fields causing hallucination · retry loops on absent data · averaging conflicting statistics.

    Ask about this article

    Get answers grounded in this post. AI-generated — based on this article, and may be imperfect.

    Free: CCA Foundations cheat-sheet (PDF)

    The domains, the 3 universal rules, core concepts, and exam-day shortcuts — one page. Enter your email and it's yours, plus my weekly AI-architecture notes.

    No spam. Unsubscribe any time.

    Scaled AI Weekly

    Enjoyed this? Get more like it every Monday.

    Real architecture decisions, LLMOps patterns that survive production, and engineering leadership advice — from 12+ years of building at enterprise scale. Free. No spam. Unsubscribe anytime.

    Join engineers building production AI systems

    Comments