What problem does it solve? Reviewing AEM Edge Delivery Services code manually is error-prone: reviewers must remember project-specific rules like editing only src/ sources, using @/* import aliases, sanitizing innerHTML with DOMPurify, and scoping CSS selectors. This Skill automates that review with structured checklists, blocking-issue detection, and visual screenshot validation before code reaches a pull request or merge. ## Core Features & Use Cases - Dual review modes: Run a self-review of uncommitted working-directory changes before committing, or review an existing GitHub pull request by number or URL. - Project-specific rule enforcement: Validates TypeScript block signatures, @/* import aliases, null-guarded querySelector calls, replaceChildren() DOM rebuilds, DOMPurify sanitization, design-token CSS variables, and range-syntax media queries. - Visual validation with Playwright: Captures desktop (1200px), tablet (768px), and mobile (375px) screenshots of preview URLs to detect layout, responsive, and rendering issues. - Actionable PR feedback: Posts review comments with GitHub Suggestions via the Pull Request Reviews API so authors can apply fixes with one click. - Use Case: After implementing a new hero block in src/blocks/hero/, invoke the skill to verify linting passes, no generated files were edited, XSS risks are sanitized, and the block renders correctly across viewports before opening the PR. ## Quick Start Ask the AI to run a code review on your uncommitted changes or on a specific pull request, for example by saying: review my current working directory changes against the AEM EDS coding standards and capture screenshots of the preview page.