design-review-workflow

Pass

Audited by Gen Agent Trust Hub on Oct 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses node:child_process functions (spawnSync, execFile) to interact with essential development and platform CLI tools.
  • scripts/check-signoff.mjs and scripts/new-review.mjs execute git commands (diff, rev-parse) to handle versioning, identify changes since sign-off, and track review commits.
  • templates/scripts/design-review/device-settings.ts and examples/capture-sweep.ts construct and run commands for adb (Android), xcrun simctl (iOS), sips (macOS image processing), and maestro (E2E testing) to automate UI capture.
  • Security is maintained by deriving command arguments from local, committed configuration files and applying regex validation to sensitive fields like device UDIDs and serial numbers.
  • [DYNAMIC_EXECUTION]: The scripts/check-lib.mjs library provides an importPackage function that uses the import() expression to load Node.js modules from a local, repo-relative path (typically .skill-tooling/). This mechanism is intended for using pinned dependencies without requiring global installation.
  • [SAFE]: No evidence of data exfiltration, hardcoded credentials, prompt injection, or unauthorized network operations was found. The skill operates entirely within the local development environment using standard tools and configuration formats.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 18, 2026, 11:22 AM