hud-manual-testing_skill

Overview

This skill provides a manual testing workflow for the Claude HUD app to verify core functionality after changes. It automates a full reset, launches the app, and walks a tester through a concise verification checklist to confirm onboarding, session tracking, hooks, and state transitions work correctly.

How this skill works

The workflow runs a reset script to create a clean first-run environment, rebuilds the app, and launches it. It then guides the tester through onboarding and a core checklist: add/remove projects, detect Claude sessions, observe state transitions, and inspect lock/session files. Debug commands are included for live logs, session state inspection, and process checks.

When to use it

• After implementing significant features
• After fixing state detection or session-tracking bugs
• Before commits that touch onboarding or hook registration
• When asked to “test the app” or “verify the app works”
• After changes to native bindings or build artifacts

Best practices

• Always start with the full reset when testing first-run or hook installation flows
• Record steps and observed states for any failing item to reproduce reliably
• Use the debug commands to capture logs before restarting or resetting
• Perform quick restarts for iterative development to save time when state persistence is desired
• Confirm cleanup of lock files and orphaned PIDs before concluding tests

Example use cases

• Verify first-run onboarding and hook installation after refactoring the hook installer
• Test session detection and state transitions after modifying the HUD state tracker
• Reproduce and diagnose a stuck Working/Waiting state using lock and PID checks
• Validate UniFFI binding updates by rebuilding and confirming the Swift bindings work
• Quickly smoke-test the app after merging platform-specific fixes with a restart script