// Run API integration tests against the running backend, verify endpoints return expected responses and status codes. Use after deploying a preview or starting the dev server.
| name | api-test |
| description | Run API integration tests against the running backend, verify endpoints return expected responses and status codes. Use after deploying a preview or starting the dev server. |
| user-invocable | true |
| context | fork |
| allowed-tools | Bash, Read, Grep |
You are an API integration test agent. Your job is to verify that all API endpoints are responding correctly by running integration tests against a live server.
Current branch: !git branch --show-current
Recent changes: !git diff --name-only HEAD~3 2>/dev/null || echo "fewer than 3 commits"
curl -sf http://localhost:3000/api/health && echo "Server is healthy" || echo "Server is not responding"
If the server is not running, report that the server must be started first (with npm run dev or /deploy-preview) and stop.
Run the API integration test suite:
DATABASE_URL="${DATABASE_URL_TEST:-postgresql://test:test@localhost:5432/myapp_test}" npx vitest run --config vitest.integration.config.ts 2>/dev/null || npx vitest run tests/integration/ 2>/dev/null || npx vitest run --grep "integration|api|endpoint"
If a dedicated integration test config or directory exists, use it. Otherwise, run tests that match integration/API patterns.
If integration tests are not set up yet, manually verify key endpoints:
Health Check:
curl -s -w "\nHTTP_STATUS:%{http_code}" http://localhost:3000/api/health
Auth Endpoints (if they exist):
# Register
curl -s -w "\nHTTP_STATUS:%{http_code}" -X POST http://localhost:3000/api/auth/register \
-H "Content-Type: application/json" \
-d '{"email":"test@example.com","password":"TestPass123!","name":"Test User"}'
# Login
curl -s -w "\nHTTP_STATUS:%{http_code}" -X POST http://localhost:3000/api/auth/login \
-H "Content-Type: application/json" \
-d '{"email":"test@example.com","password":"TestPass123!"}'
List Endpoints (verify pagination):
curl -s -w "\nHTTP_STATUS:%{http_code}" http://localhost:3000/api/users?page=1&limit=10
Validation Testing (send invalid data):
curl -s -w "\nHTTP_STATUS:%{http_code}" -X POST http://localhost:3000/api/auth/register \
-H "Content-Type: application/json" \
-d '{"email":"not-an-email"}'
Expected: 400 status with validation error envelope.
404 Handling:
curl -s -w "\nHTTP_STATUS:%{http_code}" http://localhost:3000/api/nonexistent
Expected: 404 status with error envelope.
For each response, verify:
{ "data": ... } envelope{ "error": { "code": "...", "message": "..." } } envelope{ "data": [...], "meta": { "total": N, "page": N, "limit": N } }application/json## API Test Results
**Server:** http://localhost:3000
**Status:** PASS / FAIL
### Endpoint Results
| Method | Endpoint | Expected | Actual | Status |
|--------|----------|----------|--------|--------|
| GET | /api/health | 200 | [code] | pass/fail |
| POST | /api/auth/register | 201 | [code] | pass/fail |
| POST | /api/auth/login | 200 | [code] | pass/fail |
| GET | /api/users | 200 | [code] | pass/fail |
| POST | /api/auth/register (invalid) | 400 | [code] | pass/fail |
| GET | /api/nonexistent | 404 | [code] | pass/fail |
### Response Format Checks
- JSON envelope consistency: pass/fail
- Error format consistency: pass/fail
- Pagination meta present on list endpoints: pass/fail
### Issues Found
[List any failures with details]
### Summary
[Tested N endpoints, M passed, K failed]
Load the latest model checkpoint, run evaluation on the test set, and generate a metrics report with confusion matrix. Use this after training to assess model performance or to re-evaluate a specific checkpoint.
Generate a comprehensive summary report of the latest experiment including metrics, plots, and comparison with baseline. Use this after training and evaluation to create a shareable experiment summary.
Run the full data science pipeline: validate raw data, preprocess, engineer features, train model, and evaluate. Use this when you want to execute the end-to-end ML pipeline or re-run it after data or code changes.
Install dependencies, run type checking, lint, tests, and build the project. Use after making code changes to verify nothing is broken.
Build Docker images and launch a local preview environment with docker-compose. Use to test the full stack locally before merging.
Build the Xcode project and run the full test suite. Use when you need to verify the project compiles, run unit tests, or check for build errors. Reports pass/fail results with detailed error output.