The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Model Context Protocol (MCP) tools, specifically plugin-playwright-playwright and cursor-ide-browser, to interact with the web browser. It executes commands for navigation (browser_navigate), viewport adjustments (browser_resize), and image capture (browser_take_screenshot) to validate the application's UI.
[PROMPT_INJECTION]: The skill processes untrusted external data from project files such as DESIGN_BRIEF.md and various source code files, creating a surface for indirect instructions. Ingestion points: The skill reads DESIGN_BRIEF.md from the project root or the .design/ directory, and analyzes components and style files. Boundary markers: No specific delimiters or safety instructions are used to isolate the data content from the agent's instructions. Capability inventory: The agent can navigate a browser, capture screenshots, and write design review documentation to the file system. Sanitization: No explicit sanitization or filtering of the ingested file content is performed.

design-review