The Agent Skills Directory

[Privilege Escalation] (HIGH): The documentation suggests 'sudo gem install xcodeproj', which promotes unsafe administrative privilege escalation on the host system.
[Command Execution] (HIGH): The provided usage examples interpolate a '' directly into a shell command without escaping or sanitization, allowing for arbitrary command execution via maliciously named files (e.g., '$(touch exploit).swift'). Additionally, the 'Quick Reference' provides a complex shell pipeline involving command substitution to locate and execute its own scripts, which is an unsafe execution pattern.
[Indirect Prompt Injection] (HIGH): The skill ingests untrusted file paths from user-influenced agent tasks and processes them through shell commands and Ruby logic without adequate boundary markers or sanitization, exposing a high-risk capability (modifying Xcode build configurations) to malicious input.
[Unverifiable Dependencies] (MEDIUM): The skill relies on the 'xcodeproj' gem from an external registry (RubyGems) without version pinning or integrity checks, creating a supply chain risk.
[Prompt Injection] (LOW): The use of 'MANDATORY' and 'REQUIRED' instructions is designed to override agent reasoning to ensure specific script execution, mimicking behavior-overriding injection patterns.

add-to-xcode