Skip to content

MaarkssDN/Birthub-360 / .agents/skills/appium-testing

appium-testing

Verified

Use whenever one mobile UI test suite must run on iOS and Android (native or React Native apps), or an Appium suite fails, flakes or breaks after an upgrade: XCUITest, UiAutomator2 and Espresso drivers, capabilities, clients in JS, Python, Java, Ruby or .NET, simulators, emulators and device farms.

evals/clean-machine-cannot-run-suite/criteria.json

View on GitHub
{
  "context": "In the current major the per-platform automation backends are not bundled with the server: each is a separately installed, server-managed driver, added by short name through the server's own driver subcommand and listed back by it. Predicted baseline failure: the agent reads 'could not find a driver' as a missing npm package or a missing SDK and adds `npm install -g appium-uiautomator2-driver` / an `appium-xcuitest-driver` devDependency, or piles on more Android SDK and Xcode environment plumbing, or writes 'ensure the drivers are installed' into the onboarding doc without a command. Every one of those leaves a clean machine in exactly the state described. The secondary prediction is on the authored files: capabilities written without the vendor namespace (rejected outright by this major), a hard-coded simulator or emulator model, and an onboarding doc that still tells the developer to keep a server running in a second terminal even though the runner can own that lifecycle. The failure is invisible to a casual reader because the script grows several plausible install lines and still exits 0, and the new config file looks like every capability example written before the major bump.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Platform backends installed through the server's own driver installer",
      "max_score": 32,
      "description": "The script installs the automation backends by name through the server's driver subcommand - `appium driver install uiautomator2` and `appium driver install xcuitest`. Installing them as npm packages (globally or as devDependencies), or not installing them at all, scores zero: that is exactly the state the machines are already in. Installing only the Android backend scores at most half - the task asks for machines prepared for both."
    },
    {
      "name": "Readiness check queries the server's driver inventory",
      "max_score": 16,
      "description": "Verification asks the server what it has registered - `appium driver list` (installed-only form is fine) - and exits non-zero naming the missing backend. A check limited to `appium --version`, `which appium`, `node --version` or a port probe scores at most half: all of those already pass on the broken machines. No verification step scores zero."
    },
    {
      "name": "iOS-only work gated to macOS",
      "max_score": 12,
      "description": "The XCUITest backend and the Xcode prerequisites are installed only on macOS (uname / OSTYPE guard or equivalent) and the Linux path completes successfully with Android alone. Installing the iOS backend unconditionally scores at most half. A script that hard-fails on Linux, or demands Xcode everywhere, scores zero."
    },
    {
      "name": "Authored capabilities carry the vendor prefix",
      "max_score": 18,
      "description": "In `wdio.conf.js`, every non-standard capability is namespaced - `appium:automationName`, `appium:deviceName`, `appium:app` (or `appium:appPackage` / `appium:appActivity`) - with `platformName` left bare. A capability object with no prefixes scores zero: this major rejects the session outright, so the delivered suite never runs. Prefixing some but not all scores at most half."
    },
    {
      "name": "Smoke test locates elements by the identifiers the app already ships",
      "max_score": 12,
      "description": "`test/specs/smoke.spec.js` uses the accessibility identifiers named in the notes (`email-field`, `password-field`, `sign-in`, `account-home`) through the accessibility-id strategy, and asserts the account screen is reached. Inventing XPath over the view hierarchy, or locating by visible label text, scores zero. A test that drives the flow but asserts nothing scores zero."
    },
    {
      "name": "Server lifecycle owned by the runner, not a second terminal",
      "max_score": 10,
      "description": "The config registers the runner's automation-server service so the server starts and stops around the run, and the onboarding doc drops the keep-a-terminal-open step. A `pretest` script backgrounding the server scores at most half - it removes the manual step but hand-rolls readiness. Leaving the second-terminal instruction in place scores zero."
    },
    {
      "name": "MUST NOT pin a specific device model",
      "max_score": 10,
      "description": "MUST NOT. Anti-pattern row \"Hard-coded device name ('iPhone 15')\". The committed capability set names a generic logical device ('Android Emulator' and equivalent), not a phone model or a hand-created AVD name, and does not pin an OS version. Any concrete model or version literal scores zero - it breaks on the first machine that does not happen to have it, which is the whole failure mode the team is trying to escape."
    },
    {
      "name": "Onboarding doc matches what was delivered",
      "max_score": 8,
      "description": "`docs/onboarding.md` describes the reworked flow: what the script now does, the macOS-only steps marked as such, and a command that actually runs the smoke test. A doc left describing the old flow scores zero; one that says 'run the script' while still listing stale manual steps scores at most half."
    },
    {
      "name": "Web suite untouched",
      "max_score": 5,
      "description": "MUST NOT modify e2e/login.spec.js. It is the web suite and out of scope; editing it scores zero here."
    }
  ]
}
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall