# Symptom-to-Architecture-Cause Mapping Diagnosing an agent's production symptom by mapping it to the architectural layer that actually caused it. ## Why the CCAR-P exam tests this Operational enablement means fixing the right layer, so the exam checks whether you can trace a visible failure to its true architectural cause instead of patching the surface. ## Definition **Symptom-to-architecture-cause mapping** is the diagnostic skill of taking a production complaint about an agent and identifying which part of the architecture is responsible — because the fix lives at the cause, not the symptom. Retrying, re-prompting, or adding examples wastes effort when the real defect is structural. Two mappings recur. When an agent **calls the wrong tool** or routes to the wrong sub-agent, the cause is almost always a weak or ambiguous **tool description** — the model chose based on the metadata you gave it, so the fix is clearer tool names, descriptions, and parameter docs, not a bigger model. When an agent is **confidently wrong** — fluent, assured, and factually off — the cause is usually **context**: relevant information was **dropped, truncated, or summarized away** before the model saw it, so it reasoned correctly over missing inputs. Other pairings follow the same discipline: repeated identical failures point at a deterministic prompt or schema defect; nondeterministic flakiness points at retrieval, ordering, or temperature. The exam-correct move is to name the layer — tool metadata, context assembly, retrieval, permissions — and fix there, rather than treating every symptom as a prompt-tuning problem. ## Exam trap "Wrong tool chosen" is a tool-description problem, and "confident but wrong" is a dropped-or-summarized-context problem — patching the prompt or swapping models when the cause is metadata or context assembly is the classic wrong answer. ## Commonly confused with - **ai-code-verification-checklist** — The checklist catches a bad output before merge; symptom-to-cause explains why the agent produced it, so you fix the architecture rather than re-reviewing forever. - **golden-paths** — Golden paths prevent whole classes of failure by design; symptom-to-cause is the reactive diagnosis you run when a failure still reaches production. ## Related - ai-code-verification-checklist - claude-code-in-cicd - golden-paths --- Source: Claude Cert Prep — an independent, unofficial CCAR-P study resource (domain P7). Not affiliated with Anthropic.