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.

    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