agentleFS
Sn4kyGitSn4kyGit/qa-agent-blueprint/.claude/skills/webdriverio-testing/SKILL.md

webdriverio-testing

Skill2 starsChanged 7 days agoApache-2.0View on GitHub

Write or run deterministic WebdriverIO specs (web, Selenium Grid remote, mobile via Appium) through the WDIO CLI runner. Loaded by test-automation when the project's "Test assets" → Framework says webdriverio; route user requests to /test-automation. Not for Playwright suites.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run via the suite's own scripts or npx wdio run wdio.conf.ts --spec <file>.

    line 54
  2. Very specific:

    Check "Test assets" for whether a spec already covers this flow — extend, don't duplicate.

    line 73
  3. Distinctive:

    Liveness preflight first: When the flow drives devices through an MCP server (e.g.

    line 84

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

WebdriverIO — spec conventions for the test-automator

This is the WDIO counterpart to docs/rules/testing/playwright-cli.md: same discipline (determinism, TC-IDs, retry policy, evidence), expressed in WDIO 9 terms. Loaded by test-automator when "Test assets" → Framework says webdriverio. Selenium Grid and Appium are WDIO operating modes, not separate frameworks here — see references/selenium-grid.md and references/appium-mobile.md and load them when the suite's config points at a remote hub or mobile capabilities.

Spec conventions

  • Location and config per projects/<KEY>/PROJECT.md → "Test assets" (often a separate repo checkout with its own wdio.conf.ts). Products may override.
  • One describe per flow; beforeEach for login/setup, matching the patterns already proven in memory/learnings/flaky-tests.md.
  • Element access via browser, $, $$. Assertions via expect-webdriverio (await expect($('#cart')).toBeDisplayed(), await expect(browser).toHaveUrl(...)).
  • Waiting is explicit and condition-based: waitForDisplayed(), waitForEnabled(), browser.waitUntil(). browser.pause() is the documented last resort — same standing rule as Playwright's waitForTimeout — and every use needs a code comment naming the condition that cannot be awaited yet.
  • Isolate test data per run/worker (wdio parallelizes by spec file); never assert against global/shared lists.

TC-IDs

Format TC-<CAT>-<NNN>, categories from the project's/product's "TC categories". IDs go in the it title (it('TC-CHK-014: coupon applies discount', ...)) so a failure maps directly to a reportable identifier — a wrong or invented ID silently joins two cases' histories (see test-automator's test_cases[] rule).

Retries & flakes

  • Local: specFileRetries: 0 (or --spec-file-retries=0). A flake must reproduce locally to be fixable — don't mask it with retries during development.
  • CI: match the project's config (typically specFileRetries: 2). The retry exists to capture the flake's logs and screenshots, not to turn red green.
  • A spec that passes only on retry is a flake, not a pass — log it in memory/learnings/flaky-tests.md and report it as a suite-health signal.

Runner, reporters, test_cases[]

  • Run via the suite's own scripts or npx wdio run wdio.conf.ts --spec <file>; never invent a second config.
  • test_cases[] in execution_summary.json is produced from the suite's machine reporter — enable wdio-json-reporter (or read the junit XML) for the run; the fields needed are exactly the five test-automator must persist: id (the TC-ID), outcome, attempts, spec (file), and on failure the assertion text. If no machine reporter exists in the suite, say so in the evidence instead of parsing console prose.

Evidence

Same policy as everywhere: copy run artifacts (reports, failure screenshots from outputDir, Grid video/logs where available — see the Grid reference) into evidence/<KEY>/<TICKET>/<YYYYMMDD-HHMMSS>/ per .claude/rules/core/evidence-policy.md, including execution_summary.json with test_cases[], and keep progress.md live while running.

Before writing a new spec

Check "Test assets" for whether a spec already covers this flow — extend, don't duplicate. Check memory/learnings/selector-failures.md and flaky-tests.md before choosing selectors or wait strategies. Never write a spec blind: run /recon first — WDIO suites use the probe variant recon.wdio.spec.js (same recon.json contract as the Playwright probe; console/network capture needs a devtools session and stays empty otherwise). Scaffolding a brand-new WDIO suite from nothing is not covered here (scaffold-suite is Playwright-only) — flag it to the user instead of improvising a project layout.

Before a long or unattended run

  • Liveness preflight first. When the flow drives devices through an MCP server (e.g. a wdio/Appium MCP), one cheap probe before the suite — a server down at the harness level (CONNECTION_CLOSED et al.) cannot be retried into health (field report 2026-09: 11 cases lost overnight to exactly that). Report BLOCKED with the reason and stop; reconnecting (/mcp, restart) is a human step. (The general preflight norm — preflight.sh <KEY> plus the pre-tool guard — lives in docs/rules/testing/preflight.md and test-automation; this bullet is the device-specific addendum.) Resource contention counts too: one Appium/simulator pipeline per machine at a time — two in parallel crash-loop WebDriverAgent.
  • Environment second. A clean tenant and a fresh redeploy beat a midnight run against clutter — leftover test data degrades the environment until unrelated cases fail (see "Test-data hygiene" in test-automation).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.