swiftui-skills
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The
setup.shscript executes local shell commands such asfind,cp, andxcodebuildto locate and extract documentation from the Xcode application bundle. These operations are limited to the local filesystem and do not involve network activity, privilege escalation, or persistence. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from Markdown files extracted to the
docs/directory. While these files influence the agent's behavior and code generation, the risk is minimized as the source is the official Xcode installation residing on the user's local system. - Ingestion points: Markdown files in the
docs/folder (e.g.,docs/SwiftUI-WebKit-Integration.md). - Boundary markers: The system prompt explicitly defines the
/docsdirectory as the source of truth, establishing a clear boundary for factual claims. - Capability inventory: The agent has the capability to generate and review Swift code based on the documentation, as well as accessing the user's Xcode environment as indicated in the metadata.
- Sanitization: No explicit sanitization of the documentation is performed, as the content is sourced from a trusted local application bundle.
Audit Metadata