swiftui-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides templates for UI components such as TextField and TextEditor that accept user-provided text. This establishes a surface for indirect prompt injection if the data captured is later processed by an LLM.
  • Ingestion points: Found in EditView (SKILL.md), ProfileForm (references/architecture-patterns.md), and PollView (references/design-polish.md).
  • Boundary markers: Not specified in the architectural examples.
  • Capability inventory: Includes network operations via URLSession and APIClient and clipboard manipulation via UIPasteboard (references/platform-and-sharing.md).
  • Sanitization: Not present in the provided UI pattern examples.
  • [EXTERNAL_DOWNLOADS]: The documentation contains multiple links to sosumi.ai for referencing SwiftUI and Core Haptics API documentation. These are used solely for informational purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 03:44 PM
Security Audit — agent-trust-hub — swiftui-patterns