What problem does it solve? Getting plan approval usually means pasting a plan into chat and waiting for a typed yes, which loses precision when feedback refers to specific sections. This Skill serves local plan and HTML artifacts in a browser where the reviewer annotates exact elements, chats, and returns an approve or request-changes verdict as structured JSON. ## Core Features & Use Cases - Browser Annotation Review: Renders Markdown plans (including Mermaid diagrams) or HTML artifacts on a local loopback server with an annotation layer anchored to page elements. - Blocking Feedback Loop: The await CLI command blocks until the human responds, returning annotations, chat messages, and verdicts as JSON. - Live-Reload Revision Cycle: Edit the artifact file and the canvas refreshes automatically, keeping the review loop going until approval. - Use Case: After writing .claude/plans/feature.plan.md, open it in the canvas, block on await, receive anchored feedback like "Split this into two phases" on a specific heading, revise, and get an approve verdict to start implementation. ## Quick Start Open my plan file in the browser canvas and wait for my annotated feedback and approval verdict.
