← ClaudeAtlas

testrun-forgemasterlisted

Writes and runs automated test code (XCTest, XCUITest, Espresso, JUnit, Jest, Vitest, Playwright, pytest, go test, etc.) strictly from test case specs already defined in docs/test-cases/, then reports results in plain language. Use when the user wants tests written as code, executed, or verified: "run the smoke tests", "write the UI test for SSO", "does login still pass after the refactor", "run the unit tests for checkout", or a bare "make tests for X" / "test this feature" when no test case spec is being requested. Only writes tests, runs them, and reports; never modifies app code. If no spec exists for the feature, it recommends running testcase-artisan first. Do NOT use for writing or updating test case specs or scenario lists - that is testcase-artisan's job.
zhuhrymhd/quality-atelier · ★ 0 · Testing & QA · score 67
Install: claude install-skill zhuhrymhd/quality-atelier
# Testrun Forgemaster Turns test case specs into test code, runs it, and reports what happened. Read this whole file before doing anything - the steps depend on each other. ## Boundary This skill does three things and nothing else: 1. Writes test code that implements test cases already defined in `docs/test-cases/`. 2. Runs tests. 3. Reports results, and may suggest fixes. It must never: - Modify application code, including adding accessibility identifiers, test hooks, or dependency injection seams. If a test needs one, suggest it in the report. - Weaken an assertion, add a retry, or raise a timeout just to make a failing test pass. A failing test is information, not a problem to hide. - Write to `docs/test-cases/` or `.quality-atelier.json` beyond its own section. Test case specs belong to testcase-artisan. - Edit, rename, or delete test code it didn't write. Test functions without a `Covers:` comment are the user's and stay untouched, even if they look redundant. - Add dependencies, create test targets, or edit build files (`project.pbxproj`, `build.gradle`, `package.json`, etc.). If one is needed, ask the user. - Apply its own suggested fixes, including fixes to test code it wrote. Suggest, then stop. The user decides. It only implements what the spec says. If the user asks for a test that isn't described in `docs/test-cases/`, don't improvise one. ## Step 1: Load configuration Read `.quality-atelier.json` at the repo root and use the `testrun-forgemas