Presence answers one question
Documentation can tell you whether a workflow is offered and under what conditions. It cannot, on its own, tell you how accurately, consistently, or creatively that workflow performs. A larger capability list should never automatically produce a higher quality score.
A score needs a defined test
A meaningful result needs a task definition, inputs, configuration, evaluation criteria, run date, and observed output. Repeated runs and independent judgments may be needed where output varies or quality is subjective.
Missing evidence is not failure
Our current research is selective. “Not yet assessed” means the record does not establish the capability. It does not mean the product cannot do it. “Partial evidence” signals a source limitation and deserves extra checking.
Dates have specific meanings
The source review date says when a claim was checked during research. It is not a product release date or a date on which we ran the product. We expose that distinction so readers can judge freshness without a false impression of live testing.
What happens next
The methodology sets out task-specific evaluation requirements. Until those runs exist, the site compares attributed capabilities and caveats without star ratings or numerical performance rankings.