feat: implement comprehensive testing framework and regression suite

- Established a centralized test specification in docs/tests.md that defines test categories, cases, and a coverage checklist to ensure consistent quality and maintainability across the project.

- Implemented RegressionTestExecutor for managing automated regression tests on a live Zabbix system.

- Updated GraphQL schema and resolvers with a generic runAllRegressionTests mutation.

- Enhanced MCP integration with new operation files and detailed documentation for AI-driven automation.

- Updated README.md and How-To guides (Cookbook, Maintenance, MCP) to reflect the new testing framework and MCP capabilities.

- Verified all changes with a full Jest suite (74 tests) and live end-to-end smoketests.

docs/howtos/README.md

@@ -19,6 +19,9 @@ Discover how the permission system works, how to define permission levels using
 ### 🛠️ [Technical Maintenance](./maintenance.md)
 Guide on code generation (GraphQL Codegen), running Jest tests, and local Docker builds.
 
+### 🧪 [Test Specification](../tests.md)
+Detailed list of test cases, categories (Unit, Integration, E2E), and coverage checklist.
+
 ### 🤖 [MCP & Agent Integration](./mcp.md)
 Discover how to integrate with the Model Context Protocol (MCP) to enable LLMs and autonomous agents to interact with Zabbix efficiently.
 

docs/howtos/cookbook.md

@@ -304,6 +304,31 @@ For detailed examples of the input structures, refer to [Sample Import Templates
 
 ---
 
+## 🍳 Recipe: Running the Smoketest via MCP
+
+This recipe explains how to execute the end-to-end smoketest using the integrated MCP server. This is the fastest way to verify that your Zabbix GraphQL API is correctly connected to a Zabbix instance and all core features (Groups, Templates, Hosts) are working.
+
+### 📋 Prerequisites
+- Zabbix GraphQL API is running (`npm run start` or via Docker).
+- Integrated MCP server is configured in your environment (e.g. registered in **Junie**).
+
+### 🛠️ Step 1: Prompt the AI Agent
+If you are using **Junie**, you can trigger the smoketest with a single natural language prompt.
+
+**Prompt**:
+> "Run the Zabbix smoketest using the MCP tool. Use 'SMOKE_HOST', 'SMOKE_TEMPLATE', and 'SMOKE_GROUP' for the entity names."
+
+### 🚀 Step 2: Agent Execution
+The agent will:
+1. Identify the `RunSmoketest` tool from the MCP server.
+2. Call the tool with the provided arguments.
+3. Monitor the progress of each step (Create Template Group -> Create Template -> Create Host Group -> Create and Link Host -> Verify -> Cleanup).
+
+### ✅ Step 3: Verification
+The agent will report the success or failure of each step. You should see a final message indicating "Smoketest passed successfully".
+
+---
+
 ## 🍳 Recipe: Cloning a Template with Items
 
 This recipe guides you through cloning an existing Zabbix template, including all its items and their configurations, into a new template using the GraphQL API and MCP.
@@ -379,6 +404,10 @@ AI agents can use the following MCP tools to automate this:
 - **GetTemplates**: To fetch the source template and its hierarchical item structure.
 - **ImportTemplates**: To provision the new cloned template.
 
+#### 🤖 Prompting Junie
+You can ask **Junie** to automate the entire cloning process:
+> "Using MCP, clone the template 'Generic SNMP' to a new template named 'Custom SNMP v2'. Ensure all items are copied and dependent items have their master item keys correctly mapped."
+
 ---
 
 ## 🍳 Recipe: Setting up GraphQL MCP for AI Agents

docs/howtos/maintenance.md

@@ -37,6 +37,7 @@ We use [Jest](https://jestjs.io/) for unit and integration testing.
 #### Adding New Tests
 - **Location**: Place new test files in `src/test/` with the `.test.ts` extension.
 - **Coverage**: Ensure you cover both successful operations and error scenarios.
+- **Test Specification**: Every new test must be documented in the [Test Specification](../tests.md).
 - **Best Practice**: If you find a bug, first create a reproduction test in `src/test/` to verify the fix.
 
 ## 🔄 Updating Dependencies

docs/howtos/mcp.md

@@ -63,6 +63,23 @@ To use this integration with Claude Desktop, add the following configuration to
 #### 💡 Sample Configuration (Alternative)
 If you prefer to run the MCP server manually or via Docker Compose (as described above), you can use a HTTP-based configuration instead of the `command` execution. See [.ai/mcp/mcp.json](../../.ai/mcp/mcp.json) for a sample configuration that connects to the server running on `localhost:3000`.
 
+### 🤖 Prompting with Junie (Integrated MCP)
+
+When working with **Junie** in this repository, the MCP server is already registered as an integrated tool. You can simply ask Junie to perform tasks using the Zabbix GraphQL API by referring to the available MCP tools.
+
+#### 🍳 Example: Running a Smoketest
+To verify the system end-to-end, you can prompt Junie:
+> "Run the Zabbix smoketest using the MCP tool. Use 'JUNIE_MCP_HOST', 'JUNIE_MCP_TEMPLATE', and 'JUNIE_MCP_GROUP' as names."
+
+#### 🍳 Example: Cloning a Template
+To clone a template, you can provide a higher-level instruction:
+> "Using MCP, clone the template 'Generic SNMP' to a new template named 'Custom SNMP v2'. Ensure all items are copied and dependent items have their master item keys correctly mapped."
+
+Junie will then:
+- Use `GetTemplates` to fetch the source template structure.
+- Map the items and resolve master-dependent relationships.
+- Use `ImportTemplates` to create the new cloned template.
+
 ### Benefits of GraphQL-enabled MCP over REST
 
 Integrating via GraphQL offers significant advantages for AI agents and MCP compared to the traditional Zabbix JSON-RPC (REST-like) API: