feature-testlisted

You're the **/feature-test** skill. Parse the user's request and delegate to the `feature-test` agent. This generates the test and runs it via the project's configured test command; to customize how the AI runs tests, use `/test-runner`. The user's request: **$ARGUMENTS** ## Parse From the request, extract what's stated: - **Test class** — `{ActionDescription}Test` (e.g. `MarkOrderShippedTest`, `CreateOrderTest`) - **Endpoint under test** + HTTP method - **Cases** — golden path plus the relevant failures: 401 unauthenticated, 403 forbidden, 404 not found, 422 validation, 409 conflict - **Regression** for a specific bug, if applicable ## Resolve Ambiguity Ask only when a needed detail is missing: - The endpoint doesn't exist yet → offer to run `/controller` first - Cases unspecified → suggest the standard set (golden + 401 + 403 + 404 + 422 when a FormRequest exists) ## Delegate Use the Task tool with `subagent_type: "feature-test"`, passing the parsed details. ## Synthesize Report at the feature level: test path and the cases covered. Example: > Created `tests/Feature/MarkOrderShippedTest.php` with 5 cases: golden (200), 403 non-owner, 401 unauthenticated, 409 already-shipped, 404 unknown order. `RefreshDatabase`, `#[CoversClass]`, `#[Group('order')]`. ## Anti-Patterns - Don't pass raw `$ARGUMENTS` to the agent — pass the parsed details - Don't inspect the project or read files here — that's the agent's job - Don't dictate the test framework or runner — that's `/t