feat(mcp): add media emulation

[Yury Semikhatsky (yury-s)]

Summary

• add a default MCP tool for setting and clearing page media emulation
• add paired set-* and clear-* commands under the playwright-cli Emulation category

Fixes #42243

[github-actions]

Test results for "tests 1"

5 flaky

⚠️ [chromium-library] › library/video.spec.ts:664 › screencast › should capture full viewport `@chromium-ubuntu-22.04-node22`
⚠️ [chromium-page] › page/workers.spec.ts:191 › should attribute network activity for worker inside iframe to the iframe `@chromium-ubuntu-22.04-node22`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:257 › third party 'Partitioned;' cookies `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-library] › library/browsercontext-cookies-third-party.spec.ts:470 › top level 'Partitioned;' cookie and same origin iframe `@firefox-ubuntu-22.04-node20`
⚠️ [firefox-page] › page/page-event-request.spec.ts:181 › should return response body when Cross-Origin-Opener-Policy is set `@firefox-ubuntu-22.04-node20`

51691 passed, 1247 skipped

---

Merge workflow run.

[github-actions]

Test results for "MCP"

6 failed
❌ [chrome] › mcp/cli-route.spec.ts:38 › route with content-type @mcp-macos-latest-chrome
❌ [chromium] › mcp/webmcp.spec.ts:178 › browser_webmcp_call disambiguates duplicate tool names by frame @mcp-macos-latest-chromium
❌ [firefox] › mcp/cli-core.spec.ts:248 › click in iframe @mcp-windows-latest-firefox
❌ [firefox] › mcp/cli-devtools.spec.ts:115 › per-part commands extract individual parts @mcp-windows-latest-firefox
❌ [firefox] › mcp/cli-devtools.spec.ts:173 › response-body returns just the body @mcp-windows-latest-firefox
❌ [firefox] › mcp/cli-drag.spec.ts:19 › drag between elements @mcp-windows-latest-firefox

8545 passed, 1446 skipped

---

Merge workflow run.

[github-actions]

🟡 Five of the six failures are known flakes; one I can't call either way

Hi, I'm the Playwright bot and I took a look at the failing CI. Four of the six are the same
mcp-windows-latest-firefox shard misbehaving — all four tests flake there regularly on unrelated
SHAs. The webmcp one flaked twice elsewhere in the last day. The cli-route timeout on macOS/chrome
has no history at all, so I'm leaving it uncertain rather than calling it clean.

Details

The latest merged report lists 6
failures from test run 34655556923.
This PR adds MCP media emulation — tools/backend/emulation.ts, the set-*/clear-* Emulation CLI
commands, and the new cli-emulation.spec.ts / emulation.spec.ts tests. The only shared-code edits
are additive: a new emulation entry in browserTools, an 'emulation' string added to the CLI
Category union, and a value === null early return in formatObject. None of the failing tests
exercises that path.

Note the PR's own new tests — mcp/cli-emulation.spec.ts and mcp/emulation.spec.ts — all passed in
this run, on every engine and OS.

Pre-existing flake / infra

• [firefox] › mcp/cli-core.spec.ts:248 › click in iframe, mcp/cli-devtools.spec.ts:115 › per-part commands extract individual parts, mcp/cli-devtools.spec.ts:173 › response-body returns just the body, mcp/cli-drag.spec.ts:19 › drag between elements — all four on the same bot, mcp-windows-latest-firefox, in the same shard; every other bot passed these tests in this run. Each flips verdict across the test-results DB on SHAs this PR can't be responsible for, always on that same Windows/Firefox bot:
  • per-part commands extract individual parts — 8 firefox failures over 8 distinct SHAs, including tip-of-tree pushes (11ade07c, d54cb3e4) and PRs feat(trace-viewer): show test annotations in the trace viewer #42060, chore: remove experimental component testing #42168, fix(extension): do not treat orphaned preferences entries as an installed extension #42224, feat(mcp): add browser_emulate_media tool and --color-scheme option #42244, fix(tracing): allow restarting tracing after a failed stop #42425, chore(video): emit frames with real timestamps, drop frame number quantization #42480, chore(extension): bump version to 0.4.0 #42482, fix(mcp): time out extension connection when a token is passed #42525, fix(screencast): keep overlays above top layer elements #42642.
  • response-body returns just the body — 6 firefox failures over 6 SHAs, including pushes 1e6a5e63, 680e5ad5, 648a67c7 and PRs fix(cli): rewrite CLI command in installed skills #42137, chore(recorder): render highlights via the shared element highlight #42253, fix(chromium): restore automation launch flag #42403, and feat(cli): shut idle headless sessions down after an hour #42676 as recently as yesterday.
  • drag between elements — 3 firefox failures on pushes 9a19b13f, e4a6eeb6, b856094e plus PRs fix(trace-viewer): label filter inputs and keyboard-accessible error source #42469, fix(debug-controller): drop recorded actions when recording session ends #42458 and push 8db8731c.
  • click in iframe — 2 firefox failures, on push fcc765fd and PR fix(trace-viewer): store snapshot style text in an attribute #42405.
• [chromium] › mcp/webmcp.spec.ts:178 › browser_webmcp_call disambiguates duplicate tool names by frame (macos) — a young test (~27 runs in the DB) that has already failed twice in the last day somewhere this PR isn't involved: PR feat(chromium): roll to r1245 #42683 on mcp-macos-latest-chrome, and tip-of-tree push d1ead3ec on mcp-ubuntu-latest-chromium. The error is a tool-registration message mismatch in WebMCP frame disambiguation, which the emulation tool doesn't touch.

Uncertain

• [chrome] › mcp/cli-route.spec.ts:38 › route with content-type (macos) — a bare Test timeout of 30000ms exceeded. with no other error, and the test is otherwise spotless: 0 failures across ~728 chrome runs and none on any other engine in the DB. So there's no prior occurrence to point at, and by the flake bar that's not enough to call it pre-existing. Nothing points at this PR either — it doesn't touch routing, and the naked timeout with no assertion detail reads like macOS runner slowness. A re-run would settle it.

_{Triaged by the Playwright bot - agent run}