agentleFS
ruizricaruizrica/agent-pi/skills/qa-automation/qa-test-flows/SKILL.md

qa-test-flows

Skill274 starsChanged 8 months agoMITView on GitHub

Automated UI test flow framework using CDP + agent-device for native apps and agent-browser for web apps. Zero-framework approach — bash scripts orchestrate simulators with no Detox, Maestro, or Appium dependencies. Provides test lifecycle, screenshot capture, assertion helpers, and JSON report generation. Invoke when user says "run tests", "test the app", "smoke test", "regression test", "verify flows", "run e2e tests", or any task requiring automated UI testing.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 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.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All test flows follow a consistent bash template:

    line 48

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

qa-test-flows

Automated UI test flow framework using a dual-driver architecture: CDP for navigation/state control and agent-device for screenshots/visual assertions in native apps, plus agent-browser for web app testing. This is a zero-framework approach — bash scripts orchestrate the entire test lifecycle with no Detox, Maestro, or Appium dependencies.

Key Innovation: CDP + agent-device + agent-browser

Native mobile apps with complex gesture handlers (full-screen video players, swipe-based feeds) often make coordinate-based tapping unreliable. We solve this by connecting directly to the React Native Hermes runtime via Metro's CDP WebSocket, giving us:

  • Direct navigation control via navigationRef.current.navigate() — no touch coordinates needed
  • Runtime state inspection — check current route, user state, storage
  • Module access via Metro's __r() require — load any app module at runtime
  • Screenshots & visual verification via agent-device (best-in-class for simulators)
  • Web app testing via agent-browser when the same app has a web version

When to Use

  • Smoke testing before submission or release builds
  • Regression testing after navigation or feature changes
  • Flow verification (auth, navigation, data entry, checkout)
  • Platform parity checks (iOS vs Android)
  • Pre-PR validation of multi-screen flows
  • State persistence verification

Test Architecture

Dual Driver System

DriverPurposeUse For
CDP (WebSocket)Navigation, state queries, JS executionReact Native apps (Hermes runtime)
agent-deviceScreenshots, coordinate taps, swipes, accessibilityNative app simulators/emulators
agent-browserFull browser automation, DOM interactionWeb apps, PWAs, browser testing

Test Flow Format

All test flows follow a consistent bash template:

#!/bin/bash
source "$(dirname "$0")/../../lib/test-helpers.sh"
source "$(dirname "$0")/../../lib/cdp-helpers.sh"

TEST_NAME="my-flow"
setup_test "$TEST_NAME"

# Step 1: Navigate
step "Navigate to target screen"
cdp_navigate "TargetScreen"
sleep 2
take_screenshot "01-target-screen"
assert_screenshot "01-target-screen"

# Step 2: Interact
step "Perform action"
tap 200 400
sleep 1
take_screenshot "02-after-action"

# Step 3: Verify
step "Verify result"
route=$(cdp_get_route)
log_info "Route: $route"

teardown_test

Running Tests

Single Test

bash .pi/skills/qa-automation/qa-test-flows/flows/smoke/example-smoke.sh

All Tests

bash .pi/skills/qa-automation/qa-test-flows/run-all.sh

With Logging

bash .pi/skills/qa-automation/qa-test-flows/flows/smoke/example-smoke.sh 2>&1 | tee /tmp/smoke.log

Writing New Tests

1. Copy the template

cp .pi/skills/qa-automation/qa-test-flows/templates/new-flow.sh.template \
   .pi/skills/qa-automation/qa-test-flows/flows/my-suite/my-test.sh
chmod +x .pi/skills/qa-automation/qa-test-flows/flows/my-suite/my-test.sh

2. Edit the template

Replace CUSTOMIZE markers with your app-specific details.

3. Find coordinates

agent-device screenshot /tmp/debug.png
open /tmp/debug.png  # Measure tap targets

4. Run and verify

bash .pi/skills/qa-automation/qa-test-flows/flows/my-suite/my-test.sh
ls /tmp/qa-tests/screenshots/my-test/

Helper Libraries

test-helpers.sh

FunctionPurpose
setup_test "name"Initialize test, create directories
teardown_testReport results, cleanup
step "description"Log a numbered test step
log_pass/fail/info/warnStatus logging
take_screenshot "name"Capture screenshot
tap x yTap at coordinates
swipe x1 y1 x2 y2Swipe gesture
scroll_dir directionScroll up/down/left/right
assert_app_foregroundVerify app is running
assert_screenshot "name"Verify screenshot exists
assert_text_visible "text"Check accessibility tree
launch_appLaunch/relaunch the app
close_appTerminate the app

cdp-helpers.sh

FunctionPurpose
cdp_eval "expression"Execute JS in Hermes
cdp_eval_safe "expression"Eval with ErrorUtils suppression
cdp_navigate "Screen"Navigate to screen
cdp_navigate_tab "TabScreen"Navigate to tab
cdp_get_routeGet current route name
cdp_get_stateGet full nav state
cdp_go_backGo back
nav_explore/search/home/profileQuick tab navigation

File Structure

qa-test-flows/
├── SKILL.md                          # This file
├── lib/
│   ├── test-helpers.sh               # Core test framework
│   └── cdp-helpers.sh                # CDP interaction helpers
├── flows/
│   └── smoke/
│       └── example-smoke.sh          # Example smoke test
├── templates/
│   └── new-flow.sh.template          # Template for new tests
└── run-all.sh                        # Master test runner

Troubleshooting

ProblemSolution
agent-device click hangsCommands run in background with timeout — if still hanging, increase DEVICE_CMD_TIMEOUT
Dev console overlay covers screenCall dismiss_error_overlay from scroll-helpers, or use CDP to suppress LogBox
Accessibility tree is sparseReact Native trees are thinner than web — prefer coordinates + screenshots
CDP timeoutCheck dev server is running: curl $DEV_SERVER_HEALTH
Module ID changed after code updateDelete cache: rm $NAV_MODULE_CACHE — setup guard will re-scan

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.