test_runner

Test Runner

Use this skill to execute tests and understand why they might be failing.

Inputs

• Test Command: (Optional) Override defaults (e.g., npm test -- -t "search").
• Scope: Specific file or all tests.

Tooling Strategy

• Use run_command to execute tests.
• Use view_file to read failure logs or source code if tests fail.

Workflow

1. Execution

• Command: Run the appropriate test command for the environment (e.g., npm test, pytest, python -m unittest).
• Scope: Run all tests by default, or target a specific file if requested.

2. Analysis

• Pass: If all green, celebrate! 🟢
• Fail: If red 🔴, analyze the output:
  • StackTrace: Where exactly did it blow up?
  • Assertion: What was expected vs. what was received?
  • Diff: Visual comparison of the data mismatch.

3. Recommendation

• If tests failed, provide a summary of the failure.
• Suggest potential fixes or point to the line of code causing the regression.

4. Integrity

• Ensure that the testing process didn't leave behind artifacts (temp files, database entries) unless intended for debugging.

E2E Testing Patterns (Advanced)

When running or debugging End-to-End (E2E) tests:

• Isolation: Ensure tests run in complete isolation (dedicated test accounts/data). Do not run destructive tests against production.
• Flakiness: Debug flaky tests by checking for race conditions, unstable DOM selectors, or implicit waits rather than explicit waits.
• Tracing: Utilize artifacts (screenshots, video recordings, network traces) captured during CI/CD failures to reconstruct the test state.
• Critical Journeys: Focus E2E tests on critical user journeys rather than exhaustive edge-case testing (which belongs in unit tests).