Explanation

Understanding-oriented discussion of how b8e works and why it is built the way it is. These pages are for reading away from the tool — they discuss and contextualize rather than instruct. For steps and flags, follow the links out to the how-to guides and reference.

• About the two scanning models: static source vs live DOM — why b8e scan . is a deterministic source scan and --url / --crawl / --story are a separate live-DOM lane, and why neither subsumes the other.
• About component-to-host resolution — why the scanner must map design-system wrappers to host elements, and the four-tier resolver (declaration → registry → source-trace → graph) that does it, honesty over coverage.
• About blast-radius ranking — why findings are ordered by severity first and render fan-in second, and why reach is a strict tiebreak that never overrides severity.
• About the generated reference: the docs factory — why the reference is projected from the CLI’s own sources and guarded against drift in CI, rather than written by hand.