The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were identified in the skill files. The content serves as a legitimate technical guide for SwiftUI development.
[PROMPT_INJECTION]: The skill documents an architecture for processing untrusted web data, which is its primary purpose, but it incorporates necessary safety patterns.
Ingestion points: Web content is loaded via WebView and WebPage.load() throughout the documentation and reference files.
Boundary markers: The instructions recommend using WebPage.NavigationDeciding to restrict navigation to trusted domains (e.g., docs.example.com).
Capability inventory: The skill utilizes callJavaScript for interaction and custom URL schemes for signaling.
Sanitization: The documentation explicitly advises against string interpolation for JavaScript calls, recommending structured arguments passing to mitigate injection risks.

swiftui-webkit