Siigo MCP Server

# Test Agent Skill A specialized agent for writing and running tests with dry-run mode for the Siigo MCP server. ## Overview This skill helps you write tests that: - Use **real Siigo credentials** for authentication - Execute **read operations** against the real API - **Mock write operations** (create/update/delete) to prevent mutations ## Dry-Run Mode When `DRY_RUN=true`: - `GET` requests work normally (safe, read-only) - `POST/PUT/DELETE` are intercepted and return mock responses - Operations are logged for verification ## Available Fixtures Use these in `tests/conftest.py`: ```python # Real client for read-only tests @pytest.fixture async def siigo_client(real_credentials): client = SiigoClient(**real_credentials) yield client await client.close() # Dry-run client for mutation tests @pytest.fixture async def dry_run_client(real_credentials, dry_run_env): client = DryRunSiigoClient(**real_credentials) yield client await client.close() ``` ## Writing Tests ### Read-Only Tests (Safe) ```python @pytest.mark.asyncio async def test_list_customers(siigo_client): """Safe to run - only reads data.""" result = await siigo_client.get("/customers") assert isinstance(result, dict) assert "results" in result ``` ### Mutation Tests (Use Dry-Run) ```python @pytest.mark.asyncio async def test_create_customer(dry_run_client): """Safe - uses dry-run client, no actual creation.""" result = await dry_run_client.post("/customers", { "type": "Person", "id_type": {"code": "13"}, "identification": "123456789", "name": ["John", "Doe"], }) assert result["dry_run"] is True assert result["action"] == "create" ``` ## Running Tests ```bash # Run all tests with dry-run (safe) DRY_RUN=true pytest tests/ # Run only read-only tests (always safe) pytest tests/test_reference.py # Run specific test DRY_RUN=true pytest tests/test_customers.py::test_create_customer -v ``` ## Operations Classification ### Safe to Run (Read-Only) - `get_taxes`, `get_payment_types`, `get_document_types` - `get_warehouses`, `get_users`, `get_account_groups` - `list_customers`, `get_customer` - `list_products`, `get_product` - `list_invoices`, `get_invoice`, `get_invoice_pdf` - `list_credit_notes`, `get_credit_note` - `list_journals`, `get_journal` ### Require Dry-Run (Mutations) - `create_customer`, `update_customer`, `delete_customer` - `create_product`, `update_product`, `delete_product` - `create_invoice`, `update_invoice`, `delete_invoice` - `stamp_invoice`, `send_invoice_email`, `annul_invoice` - `create_credit_note`, `create_journal` ## Test Data Examples ### Customer ```python customer_data = { "type": "Person", # or "Company" "id_type": {"code": "13"}, # 13=Cédula, 31=NIT "identification": "1234567890", "name": ["Juan", "Pérez"], "vat_responsible": False, } ``` ### Product ```python product_data = { "code": "TEST-001", "name": "Test Product", "account_group": 1234, "type": "Product", # or "Service" "stock_control": False, } ``` ### Invoice ```python invoice_data = { "document": {"id": 12345}, "customer": {"id": "customer-guid"}, "items": [{ "code": "PROD-001", "quantity": 1, "price": 10000, "taxes": [{"id": 1}], }], "payments": [{ "id": 1, "value": 10000, "due_date": "2024-12-31", }], } ``` ## Workflow 1. **Write test** using appropriate fixture (`siigo_client` or `dry_run_client`) 2. **Run with dry-run** to verify logic without mutations 3. **Check logs** to see what would have been executed 4. **Iterate** until test passes