AddThisFeature

Interaction Test Harness

Test the app the way a person uses it, through real clicks, keys, and focus.

involved Testing & QA

What it adds

A test setup that drives menus, dialogs, forms, and navigation through genuine user interactions and asserts on what the user can see.

What your agent is told to do

5
  1. 1

    Find elements the way a user finds them — by their visible label, their role, and their accessible name — rather than by class names or internal test hooks, so a test breaks when the interface becomes unusable and not when a wrapper is renamed.

  2. 2

    Drive each interaction through both pointer and keyboard where both are supported, and assert on where focus ends up as well as on what the screen shows.

  3. 3

    Assert on the visible outcome and wait for it to appear, rather than sleeping for a fixed interval and hoping. A timer-based wait is either slow on every run or flaky on a loaded machine.

  4. 4

    Cover interruptions and repeats explicitly: a dialog closed while its request is still in flight, a submit button pressed twice, a menu reopened before its close animation finishes, a navigation triggered mid-save.

  5. 5

    Rendered appearance is covered by Visual Regression Testing and accessibility rule checking by the Accessibility Test Suite; this brief owns behaviour under interaction. Do not assert on pixel positions or repeat contrast checks here.

Edge cases it handles

7
  • Selectors tied to implementation detail pass while the interface is broken and fail while it is fine, so a test that finds a control by its internal identifier is testing the wrong thing entirely.
  • A flow that works with a mouse can be unreachable by keyboard, and a suite that only clicks will never notice — every interaction test needs to say whether it exercised pointer, keyboard, or both.
  • Fixed sleeps are the main source of flake; the harness must wait for the element or the text the user would wait for, with a bounded timeout that fails loudly rather than silently passing.
  • Repeated and interrupted actions are where real defects live — double submits, a dialog dismissed before its response returns, a rapid second click on a control that was meant to disable itself — and none of these appear in a happy-path script.
  • Asserting on an element that is in the DOM but visually hidden or covered by an overlay produces a passing test for something the user cannot reach; check visibility, not presence.
  • Tests that share state through a common fixture will pass in isolation and fail when run in a different order, so each test must set up and tear down its own data.
  • Animations and transitions make elements briefly present but not yet interactive; either disable them under test or wait for the settled state rather than the first appearance.

Definition of done

8
  • Tests locate controls by visible text, role, or accessible name, and no test depends on a class name or internal identifier.
  • Each covered flow is exercised by pointer and by keyboard, with focus position asserted at the key transitions.
  • No test contains a fixed sleep; every wait is for a visible outcome with a bounded timeout.
  • Double submission, interruption mid-request, and rapid repeat interactions each have a test.
  • Assertions confirm the element is genuinely visible and reachable, not merely present in the DOM.
  • The suite passes when tests are run in a randomised order.
  • The feature matches the existing design system.
  • No existing functionality is broken.

Related features

How it works

  1. 1

    Copy the link

    Grab the Markdown instruction URL for this feature.

  2. 2

    Give it to your AI

    Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.

  3. 3

    It inspects, then implements

    Your agent reads your existing app first, then adds the feature to fit it.

Works with your stack

These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.

Need it tighter than that? Customize the feature and tell it exactly what you're running.