Test

/test

Usage

/test                     # Auto-discover and run tests
/test --create            # Generate a test suite if none exists
/test --framework <name>  # Force a framework (jest, pytest, go, ...)
/test --coverage          # Run with coverage reporting

Description

Universal test runner. Finds the right test command for the repository — checking README, package manager config, framework config, then file patterns — runs it, and reports results.

Discovery order, the per-language config/command/pattern table, coverage flags, and how to read common environment failures: → references/discovery.md

--framework skips discovery. --create generates a suite (the only mode that delegates, to test-engineer).

What matters most

• Report the real output. Stream the runner's own output and surface each failure with its file, line, and assertion. Don't summarize a failing suite as passing, and don't paraphrase counts.
• Don't fix failing tests here. This skill runs and reports; failures go back to the developer. Fixing behavior behind a green checkmark is the failure mode to avoid.
• Distinguish "no tests" from "tests failed" from "can't run tests." Each needs a different next step, and reporting a missing framework as a test failure sends the reader the wrong way.
• Ask when several suites exist (unit vs. integration vs. e2e) instead of assuming the broadest.
• With --create, run the generated tests before claiming the suite works.

Expected Output

User: /test

🔍 Discovering test command...
✅ Found: pytest -v (from pytest.ini)

🧪 Running tests...
[runner output streams through unmodified]

❌ 2 of 5 tests failed

Failed tests:
  1. test_logout (tests/test_auth.py:12) — AssertionError: assert False == True
  2. test_create_user (tests/test_api.py:25) — KeyError: 'email'

When nothing is discoverable, report what was searched (README, package config, test files, test config) and offer --create or --framework.