worktree-ios-dev

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to perform environment setup using uv tool install from a local path and provides commands for application building, testing, and simulator management via a custom tool and xcodebuild.
  • [EXTERNAL_DOWNLOADS]: The documentation suggests the optional installation of mint via brew, which is a well-known package manager for macOS, to enhance build output formatting.
  • [COMMAND_EXECUTION]: The skill includes a wipe-derived verb that executes a recursive deletion command (rm -rf) on the project's local derived data directory.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing external project configurations and CLI arguments that are interpolated into shell commands.
  • Ingestion points: Project paths (--project) and scheme names (--scheme) provided as CLI arguments.
  • Boundary markers: Absent from the provided command templates.
  • Capability inventory: Execution of subprocesses for builds, tests, and simulator lifecycle management.
  • Sanitization: No explicit sanitization or validation of input paths is described in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 09:05 AM
Security Audit — agent-trust-hub — worktree-ios-dev