The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the subprocess module to interact with standard macOS development utilities, including xcrun simctl and idb (iOS Development Bridge). This behavior is required for the skill's primary purpose of simulator automation.
Evidence of command execution found in: scripts/simctl_create.py, scripts/gesture.py, scripts/log_monitor.py, scripts/keyboard.py, scripts/navigator.py, scripts/app_state_capture.py, scripts/privacy_manager.py, scripts/status_bar.py, scripts/clipboard.py, and scripts/simulator_selector.py.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the environment that could be influenced by a malicious application running inside the simulator.
Ingestion points: The skill reads application logs via scripts/log_monitor.py and UI hierarchy/label data via scripts/common/idb_utils.py.
Boundary markers: The skill uses structured JSON output and concise text formatting to delimit data from instructions.
Capability inventory: The skill possesses the ability to write to the local filesystem (cache), control the simulator clipboard, manage app lifecycles (install/delete), and perform UI interactions (tap/type).
Sanitization: While the scripts perform structural parsing and truncation of logs, they do not specifically sanitize natural language content found within accessibility labels or log messages for prompt injection patterns.
[DATA_EXFILTRATION]: The skill interacts with sensitive data within the simulator environment, such as the system clipboard and application logs.
scripts/clipboard.py allows reading and writing to the simulator's pasteboard.
scripts/log_monitor.py and scripts/app_state_capture.py retrieve system and application logs.
Analysis shows this data remains local to the user's machine and the agent context; no external network transmission (exfiltration) patterns were detected.

ios-simulator-skill