mirror of
https://github.com/affaan-m/ECC.git
synced 2026-08-20 22:42:22 +02:00
Gaurav Dubey
9d1ecb0754
The `.opencode/commands/*.md` frontmatter referenced agents with the Claude Code plugin namespace (`agent: everything-claude-code:<name>`), but ECC's opencode integration registers its agents unscoped in `opencode.json`'s `agent` map (`code-reviewer`, `planner`, ...), and that file's own `command` section already references them unscoped. The `everything-claude-code:` scope resolves under no opencode config (the opencode plugin package is `ecc-universal`, and inline-config agents are bare), so subtask commands like `/code-review` hard-fail with `Agent not found: everything-claude-code:code-reviewer`. Non-subtask commands fall back to the default agent and appear to work — which is why only some commands failed. Strip the `everything-claude-code:` prefix from all 30 command frontmatter agent ids so they match the registered agents, fix the MIGRATION.md example, and replace the test that enforced the broken scoped invariant with one that asserts each command agent id is a registered opencode agent (fails on the old scoped ids, passes on the fix). Fixes #2477
106 lines
2.2 KiB
Markdown
106 lines
2.2 KiB
Markdown
---
|
|
description: Generate and run E2E tests with Playwright
|
|
agent: e2e-runner
|
|
subtask: true
|
|
---
|
|
|
|
# E2E Command
|
|
|
|
Generate and run end-to-end tests using Playwright: $ARGUMENTS
|
|
|
|
## Your Task
|
|
|
|
1. **Analyze user flow** to test
|
|
2. **Create test journey** with Playwright
|
|
3. **Run tests** and capture artifacts
|
|
4. **Report results** with screenshots/videos
|
|
|
|
## Test Structure
|
|
|
|
```typescript
|
|
import { test, expect } from '@playwright/test'
|
|
|
|
test.describe('Feature: [Name]', () => {
|
|
test.beforeEach(async ({ page }) => {
|
|
// Setup: Navigate, authenticate, prepare state
|
|
})
|
|
|
|
test('should [expected behavior]', async ({ page }) => {
|
|
// Arrange: Set up test data
|
|
|
|
// Act: Perform user actions
|
|
await page.click('[data-testid="button"]')
|
|
await page.fill('[data-testid="input"]', 'value')
|
|
|
|
// Assert: Verify results
|
|
await expect(page.locator('[data-testid="result"]')).toBeVisible()
|
|
})
|
|
|
|
test.afterEach(async ({ page }, testInfo) => {
|
|
// Capture screenshot on failure
|
|
if (testInfo.status !== 'passed') {
|
|
await page.screenshot({ path: `test-results/${testInfo.title}.png` })
|
|
}
|
|
})
|
|
})
|
|
```
|
|
|
|
## Best Practices
|
|
|
|
### Selectors
|
|
- Prefer `data-testid` attributes
|
|
- Avoid CSS classes (they change)
|
|
- Use semantic selectors (roles, labels)
|
|
|
|
### Waits
|
|
- Use Playwright's auto-waiting
|
|
- Avoid `page.waitForTimeout()`
|
|
- Use `expect().toBeVisible()` for assertions
|
|
|
|
### Test Isolation
|
|
- Each test should be independent
|
|
- Clean up test data after
|
|
- Don't rely on test order
|
|
|
|
## Artifacts to Capture
|
|
|
|
- Screenshots on failure
|
|
- Videos for debugging
|
|
- Trace files for detailed analysis
|
|
- Network logs if relevant
|
|
|
|
## Test Categories
|
|
|
|
1. **Critical User Flows**
|
|
- Authentication (login, logout, signup)
|
|
- Core feature happy paths
|
|
- Payment/checkout flows
|
|
|
|
2. **Edge Cases**
|
|
- Network failures
|
|
- Invalid inputs
|
|
- Session expiry
|
|
|
|
3. **Cross-Browser**
|
|
- Chrome, Firefox, Safari
|
|
- Mobile viewports
|
|
|
|
## Report Format
|
|
|
|
```
|
|
E2E Test Results
|
|
================
|
|
PASS: Passed: X
|
|
FAIL: Failed: Y
|
|
SKIPPED: Skipped: Z
|
|
|
|
Failed Tests:
|
|
- test-name: Error message
|
|
Screenshot: path/to/screenshot.png
|
|
Video: path/to/video.webm
|
|
```
|
|
|
|
---
|
|
|
|
**TIP**: Run with `--headed` flag for debugging: `npx playwright test --headed`
|