to-ios-run

Pass

Audited by Gen Agent Trust Hub on Sep 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill provides an interface to execute arbitrary Javascript code on connected iOS devices using eval_script and call_function. Specifically, the source parameter allows the execution of an asynchronous expression directly on the device runtime.
  • [COMMAND_EXECUTION]: The skill executes system commands such as iproxy and devicectl to manage USB device connections, port forwarding, and device interaction. These commands are triggered during the connection phase for physical iOS devices.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts raw code (source) and function arguments from the agent context and executes them on the mobile device without sanitization or rewriting. This creates a vulnerability surface where a malicious actor could potentially influence the agent to generate and execute unauthorized code on the target iOS device.
  • Ingestion points: The source parameter in eval_script and the args parameter in call_function (SKILL.md).
  • Boundary markers: None identified; the skill explicitly states to call functions without rewriting the input.
  • Capability inventory: The skill uses iproxy and devicectl via shell and performs Frida/LLDB attachment to execute code on mobile devices.
  • Sanitization: Absent; the skill passes raw input to the device runtime.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 29, 2026, 03:15 AM