The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of subprocess.run and subprocess.Popen to interact with external developer tools like xcrun simctl, xcodebuild, and idb. These calls are essential for the skill's primary function of iOS automation. The scripts correctly implement these calls using list-based arguments rather than shell strings, which effectively mitigates command injection risks.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically simulator logs and UI accessibility trees via scripts/common/idb_utils.py and scripts/log_monitor.py. This provides an attack surface where an application under test could potentially influence the AI agent's behavior by placing malicious text in accessibility labels or log messages. However, the scripts act as neutral processors (counting elements, filtering logs) and do not directly execute data retrieved from the simulator.
[DATA_EXPOSURE]: Several scripts (e.g., scripts/app_state_capture.py, scripts/test_recorder.py) capture screenshots, UI hierarchies, and logs for debugging purposes. These artifacts are stored locally in the user's home directory (~/.ios-simulator-skill/) or a project-specific directory (.claude/skills/ios-simulator-skill/). This behavior is consistent with the skill's purpose as a testing and diagnostic tool.
[EXTERNAL_DOWNLOADS]: The scripts/sim_health_check.sh script provides instructions for installing well-known dependencies such as idb and the Python Pillow library via trusted package managers (Homebrew, pip). These are standard tools for the iOS development ecosystem and do not constitute a security risk.

ios-simulator-skill