Canary Browser Scripting
Guides the creation and debugging of Playwright-like browser automation scripts within the Canary QuickJS sandbox environment.
Canary Browser Scripting equips Claude with deep reference knowledge and implementation patterns for the Canary sandbox scripting API. It allows developers to seamlessly write, debug, and optimize browser automation workflows using familiar async JavaScript patterns. With built-in support for AI page inspection via snapshotForAI, semantic locator strategies, persistent named pages, file-based state persistence, and screenshot capture, this skill ensures robust automation while gracefully handling sandbox constraints such as timeouts and module restrictions.
Key Features
01Direct screenshot capture and buffer handling with saveScreenshot global utilities
02AI page inspection using snapshotForAI to generate ARIA outlines for reliable selectors
030 GitHub stars
04Playwright-like DOM manipulation, clicks, form fills, and data extraction
05Sandbox troubleshooting for CPU/wall-clock timeouts, closed pages, and missing globals
06Session persistence management across execution steps via named pages and file I/O
Use Cases
01Writing multi-step automated browser scripts for web interaction and data extraction
02Inspecting complex or dynamic web pages using semantic ARIA snapshots before taking actions
03Debugging timeouts and state-loss issues across multi-turn Canary automation sessions