java-test

The Java Test Runner is a specialized engineering utility designed to manage the lifecycle of Java backend tests within the BIC Scoreboard environment. It is primarily intended for developers and software agents tasked with validating backend changes, debugging test failures, or ensuring the integrity of the hockey scoreboard service. By interfacing directly with the Gradle build system, it abstracts away the noise of complex build outputs, focusing exclusively on the critical data required to identify regressions or functional defects.

This skill is highly effective when navigating large, multi-module projects where standard test output often contains excessive boilerplate, full stack traces, or logs that obfuscate the root cause of a failure. It employs a context-reduction strategy to minimize the token usage of LLMs, ensuring that test results remain within manageable bounds even for extensive test suites.

• Executes project-specific Java tests using the ./gradlew test wrapper, including support for targeted test class execution.

• Implements parsing logic to strip ANSI escape codes and Gradle build noise, surfacing only the necessary failure reports and test statistics.

• Provides automated summaries including total tests run, pass/fail/skip counts, and specific failure details (class.method names and truncated assertion messages).

• Supports seamless integration with continuous integration workflows and local development environments including Windows/WSL and Linux.

• Utilizes optimized execution patterns, such as the no-daemon flag, to maintain high reliability and reproducibility in containerized or ephemeral build environments.

• Users should invoke this skill during active development cycles after modifying Java logic to confirm that standard TestNG tests and Selenium integration tests remain compliant.

• The tool is constrained to Java-based testing workflows and does not replace UI-specific test runners like Vitest or Playwright, which should be handled separately.

• Expected outputs are structured as status updates (PASSED/FAILED), count summaries, and concise failure lists limited to the first five errors for rapid remediation.

• When running in resource-constrained environments, prefer the provided Task agent approach or pipe outputs through the designated scripts/parse-test-output.py utility to maintain concise interaction logs.