vitest-runner

The vitest-runner skill is a specialized agentic workflow designed for comprehensive test lifecycle management in Node.js environments using the Vitest framework. It provides a standardized interface for executing test suites, ensuring proper environment grounding, and performing failure analysis. The skill is engineered for developers and quality assurance engineers who require a predictable, reproducible testing process that integrates directly into their development flow, minimizing manual intervention while maximizing visibility into test results and code quality.

• Executes test suites in various modes: single run, watch mode, and parallel execution for optimized performance.

• Performs environment verification, including checking for Node.js versions, installed dependencies in package.json, and the presence of necessary configuration files like vitest.config.ts.

• Generates and processes coverage reports, facilitating code health assessment and identifying untested logic.

• Features an integrated failure analysis protocol that captures stdout/stderr, extracts specific assertion errors, and suggests diagnostic steps based on common failure modes like module resolution issues, syntax errors, or timeouts.

• Simplifies test discovery by automatically scanning for *.test.ts and *.spec.ts patterns, reducing the friction of finding relevant test files.

• Supports granular control through CLI options such as --grep patterns, --bail, and --update for snapshots.

• Prerequisite: Requires a Node.js project with Vitest installed. Ensure node_modules are up to date before invocation.

• Input: The skill operates on test files and configuration settings. Users should provide clear intent regarding coverage requirements or specific test subsets.

• Output: Provides formatted CLI output, summary reports, and local artifact storage in .aiwg/working/checkpoints/vitest-runner/ for persistent session tracking.

• Safety: Adheres to a non-destructive policy; it will never modify test configurations or source code without explicit user confirmation, ensuring architectural integrity.

• Escalation: Automatically escalates to the user if ambiguous test configurations are detected or if tests fail repeatedly after multiple automated retry attempts.