概要
The Upstream Evaluation skill provides a rigorous methodology for AI agents to assess external repository changes against the Navigator design philosophy. By applying five distinct decision frameworks, it categorizes changes—such as new features, API renames, or bug fixes—and dictates whether they should be adopted, skipped, or adapted. This ensures that the browser automation surface remains consistent, agent-friendly, and stable, focusing purely on the high-level decision-making process before any code is integrated. The skill produces structured summaries and schema requirements, acting as the intellectual bridge between upstream evolution and local implementation.
