Design is the beginning
An evidence hierarchy is a useful first orientation. It can indicate which study designs are generally better suited to particular questions, but it cannot decide whether a result is trustworthy or applicable on its own.
A randomized trial can still have important limitations. An observational study can still provide useful information when its design, population, and uncertainty are described accurately.
Certainty belongs to an outcome
Evidence certainty should be assessed for a specific outcome in a specific population. Collapsing every finding into one score can hide meaningful differences between outcomes and study contexts.
Keep the reasoning visible
Readers should be able to see why confidence was raised or lowered. Visible limitations and direct links to supporting sources make a synthesis easier to verify and update.