em

iOS environment bring-up for driving the em Capacitor app on BrowserStack App Automate — native (XCUITest) AND web (WKWebView) in one session. The session is created out of band of the wdio MCP (a Bash bring-up script), then driven by either wdio-MCP through a local shim or the e2e bridge/helper suite. Invoked by browser-control when the target is `ios`; not normally called directly.

Web (desktop) and Android (mobile Chrome) environment bring-up for driving the em app via the chrome-devtools MCP. Invoked by browser-control when the target is `web` or `android`; not normally called directly.

ALWAYS USE THIS SKILL to bring up a browser/app environment for a given target platform (web, android, or ios) before driving the em app. The caller (e.g. issue-repro) decides the platform and passes it in; this skill is a routing hub that delegates to the right platform sub-skill (browser-control-chrome or browser-control-ios).

Use this skill after pushing commits or when asked about CI status or to fix failing tests. It monitors GitHub Actions workflow runs for the current branch, waits for completion, returns which checks passed or failed with error details, and provides a methodology for iterating until all checks pass.

ALWAYS USE THIS SKILL when working on an issue that has "Steps to Reproduce".

ALWAYS USE THIS SKILL to run a single e2e test in the real test harness for a given target platform — web/android (Vitest + puppeteer) or iOS (WDIO + Appium + simulator) — and report pass/fail with the failing assertion. Invoked by tdd-write-failing-test and by issue-repro's validation step.

ALWAYS USE THIS SKILL immediately after reproducing a bug that has documented Steps to Reproduce, and before fixing it. Turns the reproduction into a permanent automated regression test and proves it fails for the right reason. Invoked by issue-repro after its Reproduce stage.

ALWAYS USE THIS SKILL before writing implementation code for any non-trivial change. Produces a written architectural plan grounded in the existing codebase, then critiques it — both stages, one agent, one atomic unit.

Regenerate Puppeteer image snapshots using Docker. Use this skill when Puppeteer tests fail due to missing or outdated snapshots. Only use if the UI change was intentional, matches the user’s request or if you otherwise deem it to be necessary. NEVER use this skill to mask legitimate failures. ALWAYS explain to the user why you felt you needed to update snapshots.

Use this skill when CI checks have failed. It analyzes the failure logs, identifies the specific failing test or build step, categorizes the failure type, and provides guidance on how to fix it. Use in combination with the CI Monitor skill.