test-responsive

name	test-responsive
description	Run responsive design and visual adaptation tests across 4 viewports. Use when user wants to test responsive layout, mobile design, viewport behavior, dark mode, or print styles.
allowed-tools	Bash(playwright-cli:*) Bash(npx:*) Bash(mkdir:*)

Responsive Design & Visual Adaptation

Run 18 responsive test cases across 4 parallel agents — one per viewport size. Each agent runs the full test suite at its assigned viewport.

Setup

URL="${ARGUMENTS:-http://localhost:3000}"
TIMESTAMP=$(date +%Y%m%d-%H%M%S)
OUTDIR="test-results/test-responsive/$TIMESTAMP"
mkdir -p "$OUTDIR/screenshots" "$OUTDIR/snapshots" "$OUTDIR/traces"

Before running, verify prerequisites per references/prerequisites.md.

Parallel Execution — 4 Subagents (one per viewport)

Agent 1: Mobile (-s=resp-320)

playwright-cli -s=resp-320 open "$URL"
playwright-cli -s=resp-320 resize 320 568
playwright-cli -s=resp-320 tracing-start
playwright-cli -s=resp-320 screenshot --filename="$OUTDIR/screenshots/320-initial.png"
playwright-cli -s=resp-320 snapshot --filename="$OUTDIR/snapshots/320-initial.yaml"

Agent 2: Tablet (-s=resp-768)

playwright-cli -s=resp-768 open "$URL"
playwright-cli -s=resp-768 resize 768 1024
playwright-cli -s=resp-768 tracing-start
playwright-cli -s=resp-768 screenshot --filename="$OUTDIR/screenshots/768-initial.png"
playwright-cli -s=resp-768 snapshot --filename="$OUTDIR/snapshots/768-initial.yaml"

Agent 3: Laptop (-s=resp-1024)

playwright-cli -s=resp-1024 open "$URL"
playwright-cli -s=resp-1024 resize 1024 768
playwright-cli -s=resp-1024 tracing-start
playwright-cli -s=resp-1024 screenshot --filename="$OUTDIR/screenshots/1024-initial.png"
playwright-cli -s=resp-1024 snapshot --filename="$OUTDIR/snapshots/1024-initial.yaml"

Agent 4: Desktop (-s=resp-1440)

playwright-cli -s=resp-1440 open "$URL"
playwright-cli -s=resp-1440 resize 1440 900
playwright-cli -s=resp-1440 tracing-start
playwright-cli -s=resp-1440 screenshot --filename="$OUTDIR/screenshots/1440-initial.png"
playwright-cli -s=resp-1440 snapshot --filename="$OUTDIR/snapshots/1440-initial.yaml"

Test Cases — Each Agent Runs ALL

TC	Test	How to Check
TC-R1	No horizontal overflow	run-code: document.documentElement.scrollWidth <= document.documentElement.clientWidth
TC-R2	Navigation transforms (hamburger ↔ full nav)	At 320: check for hamburger/menu button. At 1440: check for full nav links visible
TC-R3	Images/media scale within viewport	run-code to find img/video/iframe elements, check none have width > viewport width
TC-R4	Touch targets ≥48x48px (mobile) / ≥44x44px (WCAG)	run-code to measure all interactive element bounding boxes on mobile viewports
TC-R5	Font sizes ≥12px, line heights readable	run-code to sample text elements, check computed fontSize ≥12px, lineHeight ratio ≥1.4
TC-R6	No text truncation hiding critical content	run-code to find elements with overflow:hidden + text-overflow:ellipsis, flag critical content (headings, CTAs, prices)
TC-R7	Modals/dialogs fit within viewport	If modal exists: open it, run-code to check modal dimensions ≤ viewport
TC-R8	Fixed/sticky elements don't overlap content	run-code to find position:fixed/sticky elements, check they don't cover more than 20% of viewport
TC-R9	Tables responsive (scroll or stack)	run-code to find tables, check for horizontal scroll wrapper or responsive stacking
TC-R10	Videos/embeds maintain aspect ratio	run-code to find video/iframe elements, check aspect ratio is reasonable (not squished)
TC-R11	Dark mode renders correctly	run-code to emulate dark color scheme: page.emulateMedia({colorScheme:'dark'}), take screenshot
TC-R12	Dark mode text contrast maintained	In dark mode: sample text elements, verify contrast ratio ≥4.5:1
TC-R13	Dark mode images/icons visible

After running all tests at the assigned viewport, each agent captures evidence:

playwright-cli -s=<session> screenshot --filename="$OUTDIR/screenshots/<viewport>-final.png"
playwright-cli -s=<session> tracing-stop
playwright-cli -s=<session> close

Fix Loop

After execution, if any tests failed, follow the fix loop process in references/fix-loop.md. Max 3 attempts. Only re-run failed test cases.

Include fix attempt history in the report.

Report Generation

After all 4 agents complete, merge results into $OUTDIR/report.md.

Use a matrix format showing pass/fail per test case per viewport:

Test Case	320×568	768×1024	1024×768	1440×900
TC-R1	✅/❌	✅/❌	✅/❌	✅/❌
TC-R2	...	...	...	...

Cleanup

playwright-cli -s=resp-320 close 2>/dev/null
playwright-cli -s=resp-768 close 2>/dev/null
playwright-cli -s=resp-1024 close 2>/dev/null
playwright-cli -s=resp-1440 close 2>/dev/null