メニュー
Маршрутизирует старые Webwright-запросы в управляемые CloakBrowser workflows. Используй для: длинная web-задача, RPA, extraction. EN triggers: compatibility browser task, long-horizon web task.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Отлаживает browser-only failures через Chrome DevTools MCP и Playwright CLI. Используй для: консоль, сеть, runtime, layout, memory, performance, Lighthouse. EN triggers: browser debug, console, network, runtime, performance.
Маршрутизирует browser tasks только через managed Playwright CLI и Chrome DevTools MCP. Используй для: UI, визуально, скриншот, Figma, консоль, сеть, перфоманс. EN triggers: browser routing, UI validation, screenshots, visual QA, console, network.
Валидирует UI и сценарии через Playwright CLI evidence. Используй для: проверь UI, браузер, скриншот, регрессия, адаптив, бизнес-логика, визуально. EN triggers: validate UI, browser check, regression, responsive, screenshot.
Низкоуровневая browser automation через Playwright CLI. Используй для: screenshots, snapshots, headed sessions, traces, responsive, UI proof. EN triggers: Playwright CLI validation, screenshots, snapshots, traces.
Проводит visual QA для Figma, screenshots и reference images. Используй для: pixel-perfect, сравни с Figma, сравни с фото, diff. EN triggers: visual diff, pixel-perfect, compare Figma, reference image.
Интернет-ресёрч: изучи интернет, latest/current, релизы, advisories. Используй для: текущая информация. EN: web search.
| name | webwright-task |
| description | Маршрутизирует старые Webwright-запросы в управляемые CloakBrowser workflows. Используй для: длинная web-задача, RPA, extraction. EN triggers: compatibility browser task, long-horizon web task. |
This skill name is retained only as a compatibility route for existing prompts. It must never start or import Webwright. Decompose long-horizon work into managed Playwright CLI actions and approved Chrome DevTools MCP diagnostics.
Before every browser action, execute exactly
$HOME/.local/bin/cloakbrowser-cdp-health. If the command is missing or exits
nonzero, stop immediately and report NOT_PROVEN.
After a successful preflight, use only:
$HOME/.local/bin/playwright-cli; never use run-code or --filename./bin/sh -c 'exec "$HOME/.local/bin/chrome-devtools-mcp" --headless --isolated --no-usage-statistics --no-performance-crux'.Never use a Webwright runtime (including Python Webwright), stock Browser, raw
Browser, in-app Browser, browser_agent, node_repl, computer-use,
Playwright MCP, raw Playwright, bunx, npx, direct provider packages,
alternate CDP endpoints, alternate executables, alternate configs, or any
fallback. Repeat the exact health preflight before each Playwright CLI command
and before each Chrome DevTools MCP tool call.
Expected outputs:
plan.md for task intent and steps.NOT_PROVEN when the managed CloakBrowser boundary is unavailable.Use the managed Playwright CLI for browser control and screenshots. Use the approved Chrome DevTools MCP transport for runtime, network, performance, memory, or Lighthouse debugging.