skills/ulpi-io/skills/build-dmg/Gen Agent Trust Hub

build-dmg

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The helper script "build-dmg.sh" executes several macOS system utilities including "xcodebuild", "hdiutil", "PlistBuddy", and "osascript". These tools are used to automate the application packaging lifecycle, including compilation, disk image creation, and Finder window styling.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted metadata from the project environment which can influence script execution.
  • Ingestion points: The skill reads application metadata such as names and versions from "package.json", "project.yml", "VERSION", and "Info.plist" files within the project being processed.
  • Boundary markers: None. The values extracted from these files are used directly to populate shell environment variables and script templates without delimiters or warnings.
  • Capability inventory: The script performs file system operations ("rm -rf", "cp") and has the capability to execute arbitrary shell commands through the "osascript" tool (via the "do shell script" command).
  • Sanitization: No sanitization is performed on the "APP_NAME" or "VERSION" strings before they are interpolated into the "osascript" heredoc. A crafted application name containing AppleScript escape sequences could potentially execute unauthorized shell commands on the user's machine during the build process.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 06:52 PM