From agentOW
Must invoke this skill if use/match: playwright, browser_navigate, browser_snapshot, browser_screenshot, browser_click, DOM assertion, SharePoint page test, debug verification, headed browser, accessibility tree
How this skill is triggered — by the user, by Claude, or both
Slash command
/agentOW:ow-dev-playwrightThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
| Tool | Purpose |
| Tool | Purpose |
|---|---|
browser_navigate | Go to a URL |
browser_snapshot | Get accessibility tree / DOM structure (primary inspection tool) |
browser_screenshot | Save page screenshot as PNG |
browser_click | Click element by text or selector |
browser_type | Type text into an input field |
Browser profile is stored at /workspaces/.playwright-profile. The Playwright MCP server is started with --user-data-dir pointing to this directory.
browser_snapshot will show an AAD login page instead of SharePoint content. Ask the user to log in manually, then retry.After browser_navigate, the page needs time to load SPFx bundles:
browser_snapshot() to check page stateSPFx debug manifests are fetched from localhost. The debug query string redirects sp-loader to load bundles from the local dev server instead of CDN.
browser_navigate(url=<fullTestUrl>) — navigate to SharePoint page with debug paramsbrowser_snapshot() — get accessibility treebrowser_screenshot() — save visual evidencebrowser_snapshot shows "Permissions requested", the agent cannot auto-approve. Ask user to approve manually.rush start. Use ow-debuglink to get the correct URL — never hardcode ports.browser_snapshot accessibility tree to identify the correct webpart by component name or aria-label.rush start may take 10-30 seconds while webpack compiles. Poll with browser_snapshot rather than using fixed waits.i to the rush tmux pane via ow-session-send(target="agentow:rush", text="i", pressEnter=false) to invalidate, then reload.Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub kaixun96/dev.agentow --plugin agentow-copilot