# Audition > Audition tests - execute test suites and capture pass/fail results, screenshots, and artifacts for visual regression routing; invoked by orchestrate after arrange or via "/audition " - Skill: `abmach/audition` (Agent Skill) - Install (CLI): `npx skillmds@latest add abmach/audition` - Raw SKILL.md: https://api.skillmd.com/api/skills/abmach/audition/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: abmach (https://skillmd.com/u/abmach) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/abmach/audition --- # Audition Tests Execute test suites to validate implementation, capture results, and generate artifacts for analysis. This skill runs tests created by the `arrange` skill and provides outputs for the `orchestrate` skill's visual regression routing. ## Pre-flight - `{{WORKSPACE}}` = workspace root. Resolve once per session and reuse: `git rev-parse --show-toplevel`; fall back to cwd outside a git repo. - Before your first write, read `{{WORKSPACE}}/{{MAESTRO_CONFIG}}/references/conventions.md` — statuses, retries, artifact paths, and file ownership are defined there and are binding. - Working folder: `{{WORKSPACE}}` - Target folders: you only WRITE to `{{WORKSPACE}}/test-results/`. Tests under `tests/` are read-only for you — never modify test files or source code. - Required input: Test type and scope from orchestrate ## References Read reference specs on-demand when the workflow requires them — do NOT read all upfront. ### Always needed - **Repo Fingerprint (working file):** Read `{{WORKSPACE}}/knowledge/repo-fingerprint.md` — to identify the current testing stack and invocation command ### On-demand (read only when needed) - **`Testing Tech Preferences`:** Read `{{WORKSPACE}}/{{MAESTRO_CONFIG}}/references/testing-tech-preferences.md` — only if an unusual/non-standard test framework is in use - **`Testing Principles`:** Read `{{WORKSPACE}}/{{MAESTRO_CONFIG}}/references/testing-principles.md` — only if interpreting failures requires methodology context ### Cross-references For how references relate to each other, see `{{WORKSPACE}}/{{MAESTRO_CONFIG}}/references/references-map.md`. ## Validation - If required input is missing, abort with error - If test files don't exist for the specified scope, abort with error ## Core Workflow ### Phase 0: Test Discovery 1. **Read Asset Specifications:** Load the working Repo Fingerprint (and `Testing Tech Preferences` if the stack is unusual) to understand the testing stack 2. **Locate Test Files:** Find test files matching the specified scope 3. **Identify Test Framework AND Runner Command:** Determine which framework is in use and how it is actually invoked: - Inspect the project's test configuration and scripts (`package.json` scripts, `playwright.config.ts`, `pytest.ini`, etc.) - Verify which runner backs a script before trusting it — an npm `test` script may run Vitest, Jest, or Playwright - Typical invocations: `npx vitest run `, `npx jest `, `npx playwright test `, `python -m pytest `, `dotnet test` ### Phase 1: Test Execution 1. Execute the test suite with the verified command 2. Capture results: output, exit codes, and any generated artifacts 3. For E2E runs, confirm screenshot artifacts were produced where the framework config directs them (baselines → `tests/screenshots/baselines/`; runtime actuals/diffs → `test-results/`) ### Phase 2: Result Processing 1. **Analyze Test Output:** Parse test results to identify pass/fail status, failed cases, error messages, stack traces, and coverage information (if available) 2. **Resolve Artifact Paths:** Collect the concrete paths of every artifact downstream consumers may need — failure screenshots, baseline paths, diff images, logs, machine-readable summaries (e.g., `test-results/summary.json`). Never report a path you have not confirmed exists 3. **Generate Summary:** Create a structured summary of test results in the Output Format below ## Output Format The result summary is the contract between audition and orchestrate's visual regression routing. Paths must be explicit and verified. ### Success Case ``` STATUS: All tests passed Test Count: Duration: Coverage: Artifacts: summary: logs: screenshots: ``` ### Failure Case ``` STATUS: Tests failed Failed Tests: Error Details: Visual Regressions: - test: baseline: actual: diff: Artifacts: summary: logs: Recommendations: ``` For non-visual failures, omit the `Visual Regressions` block. Every listed path must exist on disk at reporting time. ## Quality Checklist Before completing the test execution: - [ ] Correct test framework and runner command identified and verified - [ ] Tests executed successfully (or failures properly captured) - [ ] Test results parsed and summarized per the Output Format - [ ] Screenshots and artifacts located and reported as verified paths - [ ] Baseline vs runtime-artifact locations distinguished correctly - [ ] No background test processes left running ## Error Handling If test execution fails: - Capture the error message and exit code - Identify if it's a configuration issue, missing dependency, or test failure - Provide clear error diagnosis and suggested fixes - Ensure partial results are still captured for analysis ## Integration Points - **Input from:** `orchestrate` skill (test scope and type) - **Output to:** `orchestrate` skill (result summary with verified artifact paths for visual regression routing) - **References:** Test files created by `arrange` skill - **Context:** Implementation from `play` skill milestones ## Execution Use the test scope from the invocation, then proceed with Phase 0: Test Discovery.