swiftui-expert-skill

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to review and improve existing SwiftUI code provided by the user, which creates a surface for indirect prompt injection. Malicious instructions embedded in the source code under review could potentially influence the agent's behavior.
  • Ingestion points: Untrusted user code enters the agent context through the review and refactoring workflows specified in SKILL.md.
  • Boundary markers: The skill does not provide explicit delimiters (e.g., XML tags or triple quotes) or specific 'ignore embedded instructions' warnings for the code it processes.
  • Capability inventory: The agent utilizing this skill typically possesses the capability to read/write files and execute commands within the developer's environment.
  • Sanitization: The skill does not instruct the agent to sanitize or validate the content of the SwiftUI code before processing it.
  • [EXTERNAL_DOWNLOADS]: The README.md file contains instructions for installing the skill using npx skills add and adding plugins to Claude Code, which involve downloading content from an external GitHub repository (github.com/avdlee/swiftui-agent-skill). These references point to the skill's own source and distribution platform.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 10:33 PM