Key points
- Explicit criteria beat vague instructions: "flag a comment only when it contradicts the code's actual behavior" versus "check that comments are accurate".
- General hedges such as "be conservative" or "only high-confidence findings" do not improve precision; specific categorical criteria do.
- False positives in one category undermine developer trust in every category.
- Criteria should say what to skip (minor style, local patterns) as well as what to report (bugs, security).
- Severity stays consistent when each level has explicit criteria and a concrete code example.
Skills the exam expects
- Write review criteria that define report-versus-skip categories instead of relying on confidence filtering.
- Temporarily disable a high-false-positive category to restore trust while improving its prompt.
- Define severity levels with explicit criteria and code examples for consistent classification.
Common misconceptions
- "Adding 'carefully' or 'only when sure' sharpens judgment." It changes nothing about what counts as an issue.
- "A confidence threshold filters out the noise." Self-reported confidence is poorly calibrated.
- "More findings means a more thorough review." Developers stop reading once most findings are noise.
Watch for
- Prompts built from adjectives rather than conditions.
- Deliberate team conventions flagged repeatedly because the skip list is missing.
- Severity ratings that drift and break downstream triage.
Practice task 4.1
56 practice items cover this task in CCA Prep Studio: flashcards, matching, scenario questions and explain-it-back exercises, each tracked toward your mastery score.