test

What This Does

Detects the project's test framework, delegates test execution to the stack's tester agent, and reports a pass/fail summary. Works with any stack by auto-detecting the available test runner.

Usage

/test                     # Run the full test suite
/test path/to/file.test   # Run a specific test file
/test --coverage          # Run with coverage reporting

How It Works

1. Detect test framework. Inspect the project for known test configuration files and scripts:

   • package.json scripts (test, vitest, jest) for JS/TS stacks
   • pytest.ini, pyproject.toml [tool.pytest], setup.cfg for Python
   • Gemfile with rspec for Ruby (bundle exec rspec)
   • go.mod for Go (go test ./...)
   • mix.exs for Elixir (mix test)
   • Cargo.toml for Rust (cargo test)
   • Falls back to the stack's default test command if detection fails.

2. Delegate to the stack's tester agent. Use the Task tool to invoke the tester sub-agent with the following context:

   • Which test framework was detected
   • Which files or directories to test (all if none specified)
   • Whether coverage was requested
   • Any previous pipeline context from .agent-pipeline/implement.md

3. Run the test suite. The tester agent executes the tests, capturing stdout and stderr. If specific files were requested, only those files are run.

4. Report results. The tester agent returns a structured summary:

   ## Test Results
   
   **Status:** PASS | FAIL
   **Total:** N tests
   **Passed:** N
   **Failed:** N
   **Skipped:** N
   **Duration:** Ns
   
   ### Failures (if any)
   - test_name: reason
   

5. Write pipeline artefact. If running as part of a sequential pipeline, write results to .agent-pipeline/test.md for the next stage.

Supported Frameworks

| Stack | Frameworks | |------------|-----------------------------------| | TypeScript | Vitest, Jest, Mocha, Playwright | | Python | pytest, unittest | | Ruby | RSpec, Minitest | | Go | go test | | Elixir | ExUnit, mix test | | Rust | cargo test |

Error Handling

• If no test framework is detected, report an error and suggest configuring one.
• If tests fail, report the failures clearly but do not attempt to fix them. That is the implementer agent's responsibility.
• If the test process crashes (non-test failure), report the crash with the full error output.