The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the eval command to dynamically set shell variables based on output from xcodebuild.
Evidence: The script in Step 4 of SKILL.md uses eval $(...) to process build settings like TARGET_BUILD_DIR and FULL_PRODUCT_NAME.
Security Concern: This pattern is susceptible to command injection if the project file (e.g., .xcodeproj) contains malicious shell instructions in its metadata. Since the agent executes these commands directly, this could lead to unauthorized system access.
[COMMAND_EXECUTION]: The skill lacks sufficient sanitization for external data ingested during the build process.
Ingestion points: Build settings from xcodebuild and values from Info.plist.
Boundary markers: Absent. The skill does not implement delimiters or warnings to ignore embedded instructions in the project data.
Capability inventory: The skill uses the Bash tool to perform complex operations, providing a significant attack surface if input is compromised.
Sanitization: No validation or escaping is performed on variables extracted from the build system before their use in shell execution.

run-simulator