swiftui-design-skill
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The installation section provides an
npxcommand to add the skill, which involves downloading and executing code from an external package registry. - [EXTERNAL_DOWNLOADS]: The 'Brand Asset Integration Protocol' in section 3 instructs the agent to use web search and computer tools to find and download external assets like SVG/PNG logos and brand fonts from various online sources.
- [INDIRECT_PROMPT_INJECTION]: The skill's workflow for searching brand guidelines and downloading assets creates a surface for processing untrusted external content.
- Ingestion points: The skill utilizes
web_searchresults and files downloaded via thecomputertool as input for generating brand specifications (SKILL.md, Section 3). - Boundary markers: There are no explicit delimiters or specific 'ignore' instructions defined to separate the ingested web data from the agent's operational instructions.
- Capability inventory: The skill has the capability to write files (
brand-spec.md), generate SwiftUI source code, and perform network operations via search tools. - Sanitization: While the skill suggests verifying technical values like hex codes and contrast ratios, it lacks mechanisms to sanitize or filter potential instructions embedded within the downloaded documentation or search results.
Audit Metadata