Review Domain
One domain, ranked findings for a named decision. Frame the candidate and decision; prefer primary systems over memory; separate fact, inference, and unknown; rank findings; stop when more reading will not change the assessment. Unknowns stay unknown; they are not findings.
Open the shared review rhythm when the assessment needs depth. After several domain reviews of the same candidate, open multi-review synthesis.
Use the dedicated review listing when one exists. Use analyze-critically for a single claim. Use synthesize-evidence-brief for multi-source research rather than a product go/no-go.