diff --git a/README.md b/README.md
index d7679ea..6622032 100644
--- a/README.md
+++ b/README.md
@@ -39,6 +39,7 @@ Built with TypeScript and designed to be serverless-ready.
   - [Completing the workflow](#completing-the-workflow)
   - [Long-running operations](#long-running-operations)
   - [Custom execution](#custom-execution)
+- [Test framework](./packages/bdd/README.md)
 - [Contributors](#contributors)
 - [Made with ❤️ at Callstack](#made-with-❤️-at-callstack)
 
@@ -426,6 +427,11 @@ If you want to handle tool execution manually, you can use `iterate` function to
 
 Have a look at how `teamwork` is implemented [here](./packages/framework/src/teamwork.ts) to understand how it works.
 
+
+### BDD Testing
+
+There's a packaged called `fabrice-ai/bdd` dedicated to unit testing - actually to Behavioral Driven Development. [Check the docs](./packages/bdd/README.md).
+
 ## Contributors
 
 <!-- ALL-CONTRIBUTORS-LIST:START -->
diff --git a/bun.lockb b/bun.lockb
index 4792988..418b1b0 100755
Binary files a/bun.lockb and b/bun.lockb differ
diff --git a/example/src/ecommerce_product_description.config.ts b/example/src/ecommerce_product_description.config.ts
new file mode 100644
index 0000000..1424c5c
--- /dev/null
+++ b/example/src/ecommerce_product_description.config.ts
@@ -0,0 +1,37 @@
+import 'dotenv/config'
+
+import { visionTool } from '@fabrice-ai/tools/vision'
+import { agent } from 'fabrice-ai/agent'
+import { workflow } from 'fabrice-ai/workflow'
+import path from 'path'
+
+const techExpert = agent({
+  description: `
+    You are skilled at extracting and describing most detailed technical information about the product from the photo.
+  `,
+  tools: {
+    visionTool,
+  },
+})
+
+const marketingManager = agent({
+  description: `
+    You are skilled at writing catchy product descriptions making customers to instantly fall in love with the product.
+    You always answer why they should buy the product, how it will make their life better, 
+    and what emotions it will evoke.
+  `,
+})
+
+export const productDescriptionWorkflow = workflow({
+  team: { techExpert, marketingManager },
+  description: `
+    Based on the picture of the product, make the product description to list it on the website.
+  `,
+  knowledge: `
+    Focus on all technical features of the product, including color, size, material, brand if possible, etc.
+    Picture is at "${path.resolve(import.meta.dirname, '../assets/example-sneakers.jpg')}". 
+  `,
+  output: `
+    Catchy product description covering all the product features.
+  `,
+})
diff --git a/example/src/ecommerce_product_description.test.ts b/example/src/ecommerce_product_description.test.ts
new file mode 100644
index 0000000..a136deb
--- /dev/null
+++ b/example/src/ecommerce_product_description.test.ts
@@ -0,0 +1,30 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+
+import { productDescriptionWorkflow } from './ecommerce_product_description.config.js'
+
+const testResults = await testwork(
+  productDescriptionWorkflow,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      techExpert: [test('0_wikipedia', 'Should use "visionTool"')],
+    },
+    workflow: [
+      test(
+        '1_photo_description',
+        'The photo shows blue pair of shoes. Make sure the description includes the color and type of the shoes'
+      ),
+    ],
+  })
+)
+
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/ecommerce_product_description.ts b/example/src/ecommerce_product_description.ts
index 2a88e66..5335dac 100644
--- a/example/src/ecommerce_product_description.ts
+++ b/example/src/ecommerce_product_description.ts
@@ -1,42 +1,10 @@
 import 'dotenv/config'
 
-import { visionTool } from '@fabrice-ai/tools/vision'
-import { agent } from 'fabrice-ai/agent'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
-import { workflow } from 'fabrice-ai/workflow'
-import path from 'path'
 
-const techExpert = agent({
-  description: `
-    You are skilled at extracting and describing most detailed technical information about the product from the photo.
-  `,
-  tools: {
-    visionTool,
-  },
-})
+import { productDescriptionWorkflow } from './ecommerce_product_description.config.js'
 
-const marketingManager = agent({
-  description: `
-    You are skilled at writing catchy product descriptions making customers to instantly fall in love with the product.
-    You always answer why they should buy the product, how it will make their life better, 
-    and what emotions it will evoke.
-  `,
-})
-
-const productDescriptionWorkflow = workflow({
-  team: { techExpert, marketingManager },
-  description: `
-    Based on the picture of the product, make the product description to list it on the website.
-  `,
-  knowledge: `
-    Focus on all technical features of the product, including color, size, material, brand if possible, etc.
-    Picture is at "${path.resolve(import.meta.dirname, '../assets/example-sneakers.jpg')}". 
-  `,
-  output: `
-    Catchy product description covering all the product features.
-  `,
-})
 const result = await teamwork(productDescriptionWorkflow)
 
 console.log(solution(result))
diff --git a/example/src/github_trending.ts b/example/src/github_trending.ts
deleted file mode 100644
index ff3aa3e..0000000
--- a/example/src/github_trending.ts
+++ /dev/null
@@ -1,52 +0,0 @@
-import 'dotenv/config'
-
-import { createFireCrawlTool } from '@fabrice-ai/tools/firecrawl'
-import { getApiKey } from '@fabrice-ai/tools/utils'
-import { agent } from 'fabrice-ai/agent'
-import { solution } from 'fabrice-ai/solution'
-import { teamwork } from 'fabrice-ai/teamwork'
-import { logger } from 'fabrice-ai/telemetry'
-import { workflow } from 'fabrice-ai/workflow'
-
-const apiKey = await getApiKey('Firecrawl.dev API Key', 'FIRECRAWL_API_KEY')
-
-const { firecrawl } = createFireCrawlTool({
-  apiKey,
-})
-
-const githubResearcher = agent({
-  description: `
-    You are skilled at browsing what's hot on Github trending page.
-  `,
-  tools: {
-    firecrawl,
-  },
-})
-
-const wrapupRedactor = agent({
-  description: `
-    Your role is to compile and summarize information.
-    You're great at creating a wrap-up reports.
-    You're famous of beautiful Markdown formatting.
-  `,
-})
-
-const wrapUpTrending = workflow({
-  team: { githubResearcher, wrapupRedactor },
-  description: `
-    Research the "https://github.com/trending/typescript" page.
-    Summarize information about 3 top projects into a comprehensive markdown report.
-    Include one sentence summary for each project.
-  `,
-  knowledge: `
-    We are preparing a report for the TypeScript community.
-  `,
-  output: `
-    Comprehensive markdown report with the top trending TypeScript projects.
-  `,
-  snapshot: logger,
-})
-
-const result = await teamwork(wrapUpTrending)
-
-console.log(solution(result))
diff --git a/example/src/github_trending_vector.config.ts b/example/src/github_trending_vector.config.ts
new file mode 100644
index 0000000..53a7286
--- /dev/null
+++ b/example/src/github_trending_vector.config.ts
@@ -0,0 +1,69 @@
+import 'dotenv/config'
+
+import { createFireCrawlTool } from '@fabrice-ai/tools/firecrawl'
+import { getApiKey } from '@fabrice-ai/tools/utils'
+import { createVectorStoreTools } from '@fabrice-ai/tools/vector'
+import { agent } from 'fabrice-ai/agent'
+import { logger } from 'fabrice-ai/telemetry'
+import { workflow } from 'fabrice-ai/workflow'
+
+import { askUser } from './tools/askUser.js'
+
+const apiKey = await getApiKey('Firecrawl.dev API Key', 'FIRECRAWL_API_KEY')
+
+const { saveDocumentInVectorStore, searchInVectorStore } = createVectorStoreTools()
+
+const { firecrawl } = createFireCrawlTool({
+  apiKey,
+})
+
+const webCrawler = agent({
+  description: `
+    You are skilled at browsing Web pages.
+    You can save the documents to Vector store for later usage.
+  `,
+  tools: {
+    firecrawl,
+    saveDocumentInVectorStore,
+  },
+})
+
+const human = agent({
+  description: `
+    You can ask user and get their answer to questions that are needed by other agents.
+  `,
+  tools: {
+    askUser,
+  },
+})
+
+const reportCompiler = agent({
+  description: `
+    You can create a comprehensive report based on the information from Vector store.
+    You're famous for beautiful Markdown formatting.
+  `,
+  tools: {
+    searchInVectorStore,
+  },
+})
+
+export const wrapUpTrending = workflow({
+  team: { webCrawler, human, reportCompiler },
+  description: `
+    Research the "https://github.com/trending/typescript" page.
+    Select 3 top projects. 
+    For each project, browse details about it on their subpages.
+    Store each page in Vector store for later usage.
+  
+    Ask user about which project he wants to learn more.
+   `,
+  knowledge: `
+    Each document in Vector store is a page from the website.
+  `,
+  output: `
+    Create a comprehensive markdown report:
+     - create a one, two sentences summary about every project.
+     - include detailed summary about the project selected by the user. 
+  `,
+  snapshot: logger,
+})
diff --git a/example/src/github_trending_vector.test.ts b/example/src/github_trending_vector.test.ts
new file mode 100644
index 0000000..100ee2d
--- /dev/null
+++ b/example/src/github_trending_vector.test.ts
@@ -0,0 +1,34 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+
+import { wrapUpTrending } from './github_trending_vector.config.js'
+
+const testResults = await testwork(
+  wrapUpTrending,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      webCrawler: [
+        test(
+          '0_webCrawler',
+          'Should use "firecrawl" to crawl Github and may store data in the vector store using "saveDocumentInVectorStore"'
+        ),
+      ],
+    },
+    workflow: [
+      test('1_check_the_list', 'Should find 3 trending projects on Github'),
+      test('2_check_the_list', 'Should ask the user for one of these projects'),
+      test('3_details', 'Should generate the report with the details of the selected project'),
+    ],
+  })
+)
+
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/github_trending_vector.ts b/example/src/github_trending_vector.ts
index b7b64b7..7076966 100644
--- a/example/src/github_trending_vector.ts
+++ b/example/src/github_trending_vector.ts
@@ -1,74 +1,9 @@
 import 'dotenv/config'
 
-import { createFireCrawlTool } from '@fabrice-ai/tools/firecrawl'
-import { getApiKey } from '@fabrice-ai/tools/utils'
-import { createVectorStoreTools } from '@fabrice-ai/tools/vector'
-import { agent } from 'fabrice-ai/agent'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
-import { logger } from 'fabrice-ai/telemetry'
-import { workflow } from 'fabrice-ai/workflow'
 
-import { askUser } from './tools/askUser.js'
-
-const apiKey = await getApiKey('Firecrawl.dev API Key', 'FIRECRAWL_API_KEY')
-
-const { saveDocumentInVectorStore, searchInVectorStore } = createVectorStoreTools()
-
-const { firecrawl } = createFireCrawlTool({
-  apiKey,
-})
-
-const webCrawler = agent({
-  description: `
-    You are skilled at browsing Web pages.
-    You can save the documents to Vector store for later usage.
-  `,
-  tools: {
-    firecrawl,
-    saveDocumentInVectorStore,
-  },
-})
-
-const human = agent({
-  description: `
-    You can ask user and get their answer to questions that are needed by other agents.
-  `,
-  tools: {
-    askUser,
-  },
-})
-
-const reportCompiler = agent({
-  description: `
-    You can create a comprehensive report based on the information from Vector store.
-    You're famous for beautiful Markdown formatting.
-  `,
-  tools: {
-    searchInVectorStore,
-  },
-})
-
-const wrapUpTrending = workflow({
-  team: { webCrawler, human, reportCompiler },
-  description: `
-    Research the "https://github.com/trending/typescript" page.
-    Select 3 top projects. 
-    For each project, browse details about it on their subpages.
-    Store each page in Vector store for later usage.
-  
-    Ask user about which project he wants to learn more.
-   `,
-  knowledge: `
-    Each document in Vector store is a page from the website.
-  `,
-  output: `
-    Create a comprehensive markdown report:
-     - create a one, two sentences summary about every project.
-     - include detailed summary about the project selected by the user. 
-  `,
-  snapshot: logger,
-})
+import { wrapUpTrending } from './github_trending_vector.config.js'
 
 const result = await teamwork(wrapUpTrending)
 
diff --git a/example/src/library_photo_to_website.config.ts b/example/src/library_photo_to_website.config.ts
new file mode 100644
index 0000000..76963ea
--- /dev/null
+++ b/example/src/library_photo_to_website.config.ts
@@ -0,0 +1,63 @@
+import 'dotenv/config'
+
+import fs from 'node:fs/promises'
+import path from 'node:path'
+
+import { createFileSystemTools } from '@fabrice-ai/tools/filesystem'
+import { visionTool } from '@fabrice-ai/tools/vision'
+import { agent } from 'fabrice-ai/agent'
+import { workflow } from 'fabrice-ai/workflow'
+
+export const workingDir = path.resolve(import.meta.dirname, '../assets/')
+
+const { saveFile, readFile, listFilesFromDirectory } = createFileSystemTools({
+  workingDir,
+})
+
+const librarian = agent({
+  description: `
+    You are skilled at scanning and identifying books in the library.
+    You can analyze the photo of the library and list all the books that you see, in details.
+  `,
+  tools: {
+    visionTool,
+  },
+})
+
+const webmaster = agent({
+  description: `
+    You are skilled at creating HTML pages. 
+    You are good at using templates for creating HTML pages.
+    You can analyze existing HTML page and replace the content with the new one.
+  `,
+  tools: {
+    saveFile,
+    readFile,
+    listFilesFromDirectory,
+  },
+})
+
+export const imagePath = path.join(workingDir, 'photo-library.jpg')
+export const outputPath = path.join(workingDir, 'library.html')
+
+await fs.rm(outputPath, { force: true })
+
+export const bookLibraryWorkflow = workflow({
+  team: { librarian, webmaster },
+  description: `
+    Analyze the photo of the library and list all the books in the library.
+    Find the best template to use for the website.
+    Use the template to create a HTML page with the list of books and save it to "${outputPath}" file.
+  `,
+  knowledge: `
+    Important information:
+    - The photo of books in the library is in the "${imagePath}" file.
+    - All available templates are in "${workingDir}" directory.
+    - You only have access to files in "${workingDir}" directory.
+    - File system operations are expensive, use them wisely. Especially saving files.
+    - Use absolute paths for tool calls.
+  `,
+  output: `
+    Valid HTML page with the list of books in the library, saved in "${outputPath}" file.
+  `,
+})
diff --git a/example/src/library_photo_to_website.test.ts b/example/src/library_photo_to_website.test.ts
new file mode 100644
index 0000000..106b0be
--- /dev/null
+++ b/example/src/library_photo_to_website.test.ts
@@ -0,0 +1,66 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+import fs from 'fs'
+
+import { bookLibraryWorkflow, outputPath, workingDir } from './library_photo_to_website.config.js'
+
+const testResults = await testwork(
+  bookLibraryWorkflow,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      librarian: [
+        test(
+          '1_vision',
+          'Librarian should use the vision tool to OCR the photo of the book library to text'
+        ),
+      ],
+      webmaster: [
+        test(
+          '2_file_operations',
+          `Webmaster is using saveFile, readFile or listFilesFromDirectory tools to operate only within the ${workingDir} directory`
+        ),
+      ],
+    },
+    workflow: [
+      test(
+        '3_search_template',
+        `Webmaster should search and MUST choose the "book_library_template.html" template from inside the ${workingDir} directory.`
+      ),
+      test(
+        '4_finalOutput',
+        'Final list of the books should be at least 5 books long and saved to the HTML file'
+      ),
+      test('5_agent_routing', `The correct agent routing is librarian -> webmaster -> webmaster`),
+      test(
+        '6_finalOutput',
+        `Final output consist "Female Masculinity" title in the ${outputPath} file`,
+        async (workflow, state) => {
+          if (!fs.existsSync(outputPath)) {
+            return {
+              passed: false,
+              reasoning: `Output file ${outputPath} does not exist`,
+              id: '6_finalOutput',
+            }
+          }
+          const htmlContent = fs.readFileSync(outputPath, 'utf-8')
+          return {
+            reasoning: "Output file includes the 'Female Masculinity' title",
+            passed: htmlContent.includes('Female Masculinity'),
+            id: '6_finalOutput',
+          }
+        }
+      ),
+    ],
+  })
+)
+
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/library_photo_to_website.ts b/example/src/library_photo_to_website.ts
index 45a8e1a..805368d 100644
--- a/example/src/library_photo_to_website.ts
+++ b/example/src/library_photo_to_website.ts
@@ -1,68 +1,9 @@
 import 'dotenv/config'
 
-import fs from 'node:fs/promises'
-import path from 'node:path'
-
-import { createFileSystemTools } from '@fabrice-ai/tools/filesystem'
-import { visionTool } from '@fabrice-ai/tools/vision'
-import { agent } from 'fabrice-ai/agent'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
-import { workflow } from 'fabrice-ai/workflow'
-
-const workingDir = path.resolve(import.meta.dirname, '../assets/')
-
-const { saveFile, readFile, listFilesFromDirectory } = createFileSystemTools({
-  workingDir,
-})
-
-const librarian = agent({
-  description: `
-    You are skilled at scanning and identifying books in the library.
-    You can analyze the photo of the library and list all the books that you see, in details.
-  `,
-  tools: {
-    visionTool,
-  },
-})
-
-const webmaster = agent({
-  description: `
-    You are skilled at creating HTML pages. 
-    You are good at using templates for creating HTML pages.
-    You can analyze existing HTML page and replace the content with the new one.
-  `,
-  tools: {
-    saveFile,
-    readFile,
-    listFilesFromDirectory,
-  },
-})
-
-const imagePath = path.join(workingDir, 'photo-library.jpg')
-const outputPath = path.join(workingDir, 'library.html')
-
-await fs.rm(outputPath, { force: true })
 
-const bookLibraryWorkflow = workflow({
-  team: { librarian, webmaster },
-  description: `
-    Analyze the photo of the library and list all the books in the library.
-    Find the best template to use for the website.
-    Use the template to create a HTML page with the list of books and save it to "${outputPath}" file.
-  `,
-  knowledge: `
-    Important information:
-    - The photo of books in the library is in the "${imagePath}" file.
-    - All available templates are in "${workingDir}" directory.
-    - You only have access to files in "${workingDir}" directory.
-    - File system operations are expensive, use them wisely. Especially saving files.
-    - Use absolute paths for tool calls.
-  `,
-  output: `
-    Valid HTML page with the list of books in the library, saved in "${outputPath}" file.
-  `,
-})
+import { bookLibraryWorkflow } from './library_photo_to_website.config.js'
 
 const result = await teamwork(bookLibraryWorkflow)
 
diff --git a/example/src/medical_survey/workflow.ts b/example/src/medical_survey.config.ts
similarity index 96%
rename from example/src/medical_survey/workflow.ts
rename to example/src/medical_survey.config.ts
index 4df57ce..18cc153 100644
--- a/example/src/medical_survey/workflow.ts
+++ b/example/src/medical_survey.config.ts
@@ -1,7 +1,7 @@
 import { agent } from 'fabrice-ai/agent'
 import { workflow } from 'fabrice-ai/workflow'
 
-import { askUser } from '../tools/askUser.js'
+import { askUser } from './tools/askUser.js'
 
 const nurse = agent({
   description: `
diff --git a/example/src/medical_survey.test.ts b/example/src/medical_survey.test.ts
new file mode 100644
index 0000000..f3a499a
--- /dev/null
+++ b/example/src/medical_survey.test.ts
@@ -0,0 +1,68 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+import { tool } from 'fabrice-ai/tool'
+import { z } from 'zod'
+
+import { preVisitNoteWorkflow } from './medical_survey.config.js'
+
+export const askUserMock = tool({
+  description: 'Tool for asking user a question',
+  parameters: z.object({
+    query: z.string().describe('The question to ask the user'),
+  }),
+  execute: async ({ query }, { provider }): Promise<string> => {
+    const response = await provider.chat({
+      messages: [
+        {
+          role: 'system',
+          content: `We are role playing - a nurse is asking a patient about their symptoms
+          and the patient is answering. The nurse will ask you a question and you should answer it.
+          Figure out something realistic! It's just a play!`,
+        },
+        {
+          role: 'user',
+          content: 'Try to answer this question in a single line: ' + query,
+        },
+      ],
+      response_format: {
+        result: z.object({
+          answer: z.string().describe('Answer to the question'),
+        }),
+      },
+    })
+    console.log(`😳 Mocked response: ${response.value.answer}\n`)
+    return Promise.resolve(response.value.answer)
+  },
+})
+
+preVisitNoteWorkflow.team['nurse'].tools = {
+  askPatient: askUserMock,
+}
+
+const testResults = await testwork(
+  preVisitNoteWorkflow,
+  suite({
+    description: 'Automated testing suite for med journey',
+    team: {
+      nurse: [
+        test(
+          '0_askPatient',
+          'Should ask question - the patient should answer. Check if the question was asked and answer provided. Do not analyze the content.'
+        ),
+      ],
+    },
+    workflow: [
+      test('1_questionare', 'Should ask up to 5 questions to the user and wait for the results'),
+      test('2_diagnosis', 'Should compile the pre-visit report the patient based on the answers'),
+    ],
+  })
+)
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/medical_survey.ts b/example/src/medical_survey.ts
index b456108..dac926c 100644
--- a/example/src/medical_survey.ts
+++ b/example/src/medical_survey.ts
@@ -3,7 +3,7 @@ import 'dotenv/config'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
 
-import { preVisitNoteWorkflow } from './medical_survey/workflow.js'
+import { preVisitNoteWorkflow } from './medical_survey.config.js'
 
 const result = await teamwork(preVisitNoteWorkflow)
 
diff --git a/example/src/medical_survey_ollama.ts b/example/src/medical_survey_ollama.ts
new file mode 100644
index 0000000..42bd6d9
--- /dev/null
+++ b/example/src/medical_survey_ollama.ts
@@ -0,0 +1,14 @@
+import { ollama } from 'fabrice-ai/providers/ollama'
+import { solution } from 'fabrice-ai/solution'
+import { teamwork } from 'fabrice-ai/teamwork'
+
+import { preVisitNoteWorkflow } from './medical_survey.config.js'
+
+const result = await teamwork({
+  ...preVisitNoteWorkflow,
+  provider: ollama({
+    model: 'llama3.1',
+  }),
+})
+
+console.log(solution(result))
diff --git a/example/src/medical_survey_server.ts b/example/src/medical_survey_server.ts
index 7079b18..33a7d03 100644
--- a/example/src/medical_survey_server.ts
+++ b/example/src/medical_survey_server.ts
@@ -16,7 +16,7 @@ import {
 } from 'fabrice-ai/tool_calls'
 import fastify, { FastifyRequest } from 'fastify'
 
-import { preVisitNoteWorkflow } from './medical_survey/workflow.js'
+import { preVisitNoteWorkflow } from './medical_survey.config.js'
 
 const server = fastify({ logger: false })
 
diff --git a/example/src/news_wrap_up.config.ts b/example/src/news_wrap_up.config.ts
new file mode 100644
index 0000000..9b0d8ce
--- /dev/null
+++ b/example/src/news_wrap_up.config.ts
@@ -0,0 +1,54 @@
+import 'dotenv/config'
+
+import { getCurrentDate } from '@fabrice-ai/tools/date'
+import { getApiKey } from '@fabrice-ai/tools/utils'
+import { createWebSearchTools } from '@fabrice-ai/tools/webSearch'
+import { agent } from 'fabrice-ai/agent'
+import { logger } from 'fabrice-ai/telemetry'
+import { workflow } from 'fabrice-ai/workflow'
+
+const apiKey = await getApiKey('Serply.io API', 'SERPLY_API_KEY')
+
+const { googleSearch } = createWebSearchTools({
+  apiKey,
+})
+
+const newsResearcher = agent({
+  description: `
+    You are skilled at searching the News over Web.
+    Your job is to get the news from the last week.
+  `,
+  tools: {
+    googleSearch,
+    getCurrentDate,
+  },
+})
+
+const newsReader = agent({
+  description: `
+    You're greatly skilled at reading and summarizing news headlines.
+    Other team members rely on you to get the gist of the news.
+    You always tries to be objective, not halucinating neither adding your own opinion.
+  `,
+})
+
+const wrapupRedactor = agent({
+  description: `
+    Your role is to wrap up the news and trends for the last week into a comprehensive report.
+    Generalization is also one of your powerfull skills, however you're not a fortune teller.
+    You're famous for precisely getting the overal picture, trends and summarizing it all.
+  `,
+})
+
+export const wrapUpTheNewsWorkflow = workflow({
+  team: { newsResearcher, newsReader, wrapupRedactor },
+  description: `
+    Research the top news and trends for the last week.
+  `,
+  output: `
+    Comprehensive markdown report with the listing including top news headlines for the last week.
+    - Include one sentence summary for each article.
+    - Include top takeaways - bulletpoints from each article.
+  `,
+  snapshot: logger,
+})
diff --git a/example/src/news_wrap_up.test.ts b/example/src/news_wrap_up.test.ts
new file mode 100644
index 0000000..4dea007
--- /dev/null
+++ b/example/src/news_wrap_up.test.ts
@@ -0,0 +1,41 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+
+import { wrapUpTheNewsWorkflow } from './news_wrap_up.config.js'
+
+const testResults = await testwork(
+  wrapUpTheNewsWorkflow,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      newsResearcher: [
+        test('1_currentDate', 'Should use "getCurrentDate" tool'),
+        test('2_googleSearch', 'Should use "googleSearch" tool'),
+      ],
+    },
+    workflow: [
+      test(
+        '3_newsReasearcher',
+        'The newsResearcher agent should take the current date and search for the news from the last week'
+      ),
+      test(
+        '4_wrapUpRedactor',
+        'Wrap up redactor should be used to compile the markdown report with the listing including top news headlines for the last week'
+      ),
+      test(
+        '5_finalOutput',
+        'Final output should be a markdown report with the listing of top news headlines'
+      ),
+    ],
+  })
+)
+
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/news_wrap_up.ts b/example/src/news_wrap_up.ts
index 71fbb94..ae2045f 100644
--- a/example/src/news_wrap_up.ts
+++ b/example/src/news_wrap_up.ts
@@ -1,59 +1,9 @@
 import 'dotenv/config'
 
-import { getCurrentDate } from '@fabrice-ai/tools/date'
-import { getApiKey } from '@fabrice-ai/tools/utils'
-import { createWebSearchTools } from '@fabrice-ai/tools/webSearch'
-import { agent } from 'fabrice-ai/agent'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
-import { logger } from 'fabrice-ai/telemetry'
-import { workflow } from 'fabrice-ai/workflow'
 
-const apiKey = await getApiKey('Serply.io API', 'SERPLY_API_KEY')
-
-const { googleSearch } = createWebSearchTools({
-  apiKey,
-})
-
-const newsResearcher = agent({
-  description: `
-    You are skilled at searching the News over Web.
-    Your job is to get the news from the last week.
-  `,
-  tools: {
-    googleSearch,
-    getCurrentDate,
-  },
-})
-
-const newsReader = agent({
-  description: `
-    You're greatly skilled at reading and summarizing news headlines.
-    Other team members rely on you to get the gist of the news.
-    You always tries to be objective, not halucinating neither adding your own opinion.
-  `,
-})
-
-const wrapupRedactor = agent({
-  description: `
-    Your role is to wrap up the news and trends for the last week into a comprehensive report.
-    Generalization is also one of your powerfull skills, however you're not a fortune teller.
-    You're famous for precisely getting the overal picture, trends and summarizing it all.
-  `,
-})
-
-const wrapUpTheNewsWorkflow = workflow({
-  team: { newsResearcher, newsReader, wrapupRedactor },
-  description: `
-    Research the top news and trends for the last week.
-  `,
-  output: `
-    Comprehensive markdown report with the listing including top news headlines for the last week.
-    - Include one sentence summary for each article.
-    - Include top takeaways - bulletpoints from each article.
-  `,
-  snapshot: logger,
-})
+import { wrapUpTheNewsWorkflow } from './news_wrap_up.config.js'
 
 const result = await teamwork(wrapUpTheNewsWorkflow)
 
diff --git a/example/src/surprise_trip.config.ts b/example/src/surprise_trip.config.ts
new file mode 100644
index 0000000..49d0d2b
--- /dev/null
+++ b/example/src/surprise_trip.config.ts
@@ -0,0 +1,79 @@
+import 'dotenv/config'
+
+import { agent } from 'fabrice-ai/agent'
+import { workflow } from 'fabrice-ai/workflow'
+
+import { lookupWikipedia } from './tools/wikipedia.js'
+
+const personalizedActivityPlanner = agent({
+  description: `
+    You are skilled at researching and finding cool things to do at the destination,
+    including activities and events that match the traveler's interests and age group.
+  `,
+})
+
+const landmarkScout = agent({
+  description: `
+    You are skilled at researching and finding interesting landmarks at the destination.
+    Your find historical landmarks, museums, and other interesting places.
+  `,
+  tools: {
+    lookupWikipedia,
+  },
+})
+
+const restaurantScout = agent({
+  description: `
+    As a food lover, you know the best spots in town for a delightful culinary experience.
+    You also have a knack for finding picturesque and entertaining locations.
+    Your find highly-rated restaurants and dining experiences at the destination,
+    and recommend scenic locations and fun activities.
+  `,
+})
+
+const itineraryCompiler = agent({
+  description: `
+    With an eye for detail, you organize all the information into a coherent and enjoyable travel plan.
+  `,
+})
+
+export const researchTripWorkflow = workflow({
+  team: {
+    personalizedActivityPlanner,
+    restaurantScout,
+    landmarkScout,
+    itineraryCompiler,
+  },
+  description: `
+    Research and find cool things to do in Wrocław, Poland.
+
+    Focus:
+      - activities and events that match the traveler's age group.
+      - highly-rated restaurants and dining experiences.
+      - landmarks with historic context.
+      - picturesque and entertaining locations.
+  `,
+  knowledge: `
+    Traveler's information:
+      - Origin: New York, USA
+      - Destination: Wrocław, Poland
+      - Age of the traveler: 30
+      - Hotel location: Hilton, Main Square, Wrocław
+      - Flight information: Flight AA123, arriving on 2023-12-15
+      - How long is the trip: 7 days
+      - Likes: history, italian food, vintage cars.
+    
+    Flights and hotels are already confirmed.
+  `,
+  output: `
+    Comprehensive day-by-day plan for the trip to Wrocław, Poland.
+    Ensure the plan includes flights, hotel information, and all planned activities and dining experiences.
+  `,
+  // Claude
+  // provider: openrouter({
+  //   model: 'anthropic/claude-3.5-haiku-20241022:beta',
+  //   structured_output: false,
+  // }),
+  // Grok
+  // provider: grok(),
+})
diff --git a/example/src/surprise_trip.test.ts b/example/src/surprise_trip.test.ts
new file mode 100644
index 0000000..ed32c4c
--- /dev/null
+++ b/example/src/surprise_trip.test.ts
@@ -0,0 +1,53 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+import { WorkflowState } from 'fabrice-ai/state'
+import { Workflow } from 'fabrice-ai/workflow'
+
+import { researchTripWorkflow } from './surprise_trip.config.js'
+
+const testResults = await testwork(
+  researchTripWorkflow,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      landmarkScout: [test('0_wikipedia', 'Should use "lookupWikipedia" tool')],
+    },
+    workflow: [
+      test(
+        '1_personalizedActivityPlanner',
+        'Should use "personalizedActivityPlanner" to "Research activities and events in Wrocław"'
+      ),
+      test(
+        '2_restaurantScout',
+        'Should use "restaurantScount" to "Research restaurants and dining experience in Wrocław"'
+      ),
+      test('3_landmarkScout', 'Should use "landmarkScout" to "Research landmarks of Wrocław"'),
+      test(
+        '5_itineraryCompiler',
+        '"itineraryCompiler" should compile all the information into a coherent travel plan'
+      ),
+      test('6_finalOutput', 'Should return a 7 days itinerary as a final output'),
+      test(
+        '7_snapshot',
+        'This is example non LLM check',
+        async (workflow: Workflow, state: WorkflowState) => {
+          return {
+            passed: true,
+            reasoning: 'This is example non LLM check',
+            id: '7_snapshot',
+          }
+        }
+      ),
+    ],
+  })
+)
+
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/surprise_trip.ts b/example/src/surprise_trip.ts
index d70e96f..d735ac4 100644
--- a/example/src/surprise_trip.ts
+++ b/example/src/surprise_trip.ts
@@ -1,86 +1,9 @@
 import 'dotenv/config'
 
-import { agent } from 'fabrice-ai/agent'
-import { grok } from 'fabrice-ai/providers/grok'
-import { openrouter } from 'fabrice-ai/providers/openrouter'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
-import { workflow } from 'fabrice-ai/workflow'
 
-import { lookupWikipedia } from './tools/wikipedia.js'
-
-const personalizedActivityPlanner = agent({
-  description: `
-    You are skilled at researching and finding cool things to do at the destination,
-    including activities and events that match the traveler's interests and age group.
-  `,
-})
-
-const landmarkScout = agent({
-  description: `
-    You are skilled at researching and finding interesting landmarks at the destination.
-    Your find historical landmarks, museums, and other interesting places.
-  `,
-  tools: {
-    lookupWikipedia,
-  },
-})
-
-const restaurantScout = agent({
-  description: `
-    As a food lover, you know the best spots in town for a delightful culinary experience.
-    You also have a knack for finding picturesque and entertaining locations.
-    Your find highly-rated restaurants and dining experiences at the destination,
-    and recommend scenic locations and fun activities.
-  `,
-})
-
-const itineraryCompiler = agent({
-  description: `
-    With an eye for detail, you organize all the information into a coherent and enjoyable travel plan.
-  `,
-})
-
-const researchTripWorkflow = workflow({
-  team: {
-    personalizedActivityPlanner,
-    restaurantScout,
-    landmarkScout,
-    itineraryCompiler,
-  },
-  description: `
-    Research and find cool things to do in Wrocław, Poland.
-
-    Focus:
-      - activities and events that match the traveler's age group.
-      - highly-rated restaurants and dining experiences.
-      - landmarks with historic context.
-      - picturesque and entertaining locations.
-  `,
-  knowledge: `
-    Traveler's information:
-      - Origin: New York, USA
-      - Destination: Wrocław, Poland
-      - Age of the traveler: 30
-      - Hotel location: Hilton, Main Square, Wrocław
-      - Flight information: Flight AA123, arriving on 2023-12-15
-      - How long is the trip: 7 days
-      - Likes: history, italian food, vintage cars.
-    
-    Flights and hotels are already confirmed.
-  `,
-  output: `
-    Comprehensive day-by-day plan for the trip to Wrocław, Poland.
-    Ensure the plan includes flights, hotel information, and all planned activities and dining experiences.
-  `,
-  // Claude
-  // provider: openrouter({
-  //   model: 'anthropic/claude-3.5-haiku-20241022:beta',
-  //   structured_output: false,
-  // }),
-  // Grok
-  // provider: grok(),
-})
+import { researchTripWorkflow } from './surprise_trip.config.js'
 
 const result = await teamwork(researchTripWorkflow)
 
diff --git a/example/src/wikipedia_vector.config.ts b/example/src/wikipedia_vector.config.ts
new file mode 100644
index 0000000..c64b902
--- /dev/null
+++ b/example/src/wikipedia_vector.config.ts
@@ -0,0 +1,50 @@
+import 'dotenv/config'
+
+import { createVectorStoreTools } from '@fabrice-ai/tools/vector'
+import { agent } from 'fabrice-ai/agent'
+import { logger } from 'fabrice-ai/telemetry'
+import { workflow } from 'fabrice-ai/workflow'
+
+import { lookupWikipedia } from './tools/wikipedia.js'
+
+const { saveDocumentInVectorStore, searchInVectorStore } = createVectorStoreTools()
+
+const wikipediaIndexer = agent({
+  description: `
+    You are skilled at reading and understanding the context of Wikipedia articles.
+    You can save information in Vector store for later use.
+    When saving articles in Vector store, you store every sentence as a separate document and 
+    you only save first 10 sentences.
+  `,
+  tools: {
+    lookupWikipedia,
+    saveDocumentInVectorStore,
+  },
+})
+
+const reportCompiler = agent({
+  description: `
+    You are skilled at compiling information from various sources into a coherent report.
+    You can search for specific sentences in Vector database.
+  `,
+  tools: {
+    searchInVectorStore,
+  },
+})
+
+export const wikipediaResearch = workflow({
+  team: { wikipediaIndexer, reportCompiler },
+  description: `
+    Find information about John III Sobieski on Wikipedia and save it in Vector store.
+    Lookup sentences related to the following topics:
+     - "Dates of reign as King of Poland"
+     - "John III education"
+  `,
+  knowledge: `
+    Each document in Vector store is a sentence from the Wikipedia article.
+  `,
+  output: `
+    List of sentences looked up for each topic. Each sentence should be in separate bullet point.
+  `,
+  snapshot: logger,
+})
diff --git a/example/src/wikipedia_vector.test.ts b/example/src/wikipedia_vector.test.ts
new file mode 100644
index 0000000..bbbb6e7
--- /dev/null
+++ b/example/src/wikipedia_vector.test.ts
@@ -0,0 +1,32 @@
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+
+import { wikipediaResearch } from './wikipedia_vector.config.js'
+
+const testResults = await testwork(
+  wikipediaResearch,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      wikipediaIndexer: [
+        test('0_wikipedia', 'Should use "wikipediaTool" and "saveDocumentInVectorStore"'),
+      ],
+    },
+    workflow: [
+      test(
+        '1_check_the_story',
+        'There should be a short report about John III Sobieski education and dates of reign as King of Poland'
+      ),
+    ],
+  })
+)
+
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
diff --git a/example/src/wikipedia_vector.ts b/example/src/wikipedia_vector.ts
index 7f3d040..5aa361b 100644
--- a/example/src/wikipedia_vector.ts
+++ b/example/src/wikipedia_vector.ts
@@ -1,55 +1,9 @@
 import 'dotenv/config'
 
-import { createVectorStoreTools } from '@fabrice-ai/tools/vector'
-import { agent } from 'fabrice-ai/agent'
 import { solution } from 'fabrice-ai/solution'
 import { teamwork } from 'fabrice-ai/teamwork'
-import { logger } from 'fabrice-ai/telemetry'
-import { workflow } from 'fabrice-ai/workflow'
 
-import { lookupWikipedia } from './tools/wikipedia.js'
-
-const { saveDocumentInVectorStore, searchInVectorStore } = createVectorStoreTools()
-
-const wikipediaIndexer = agent({
-  description: `
-    You are skilled at reading and understanding the context of Wikipedia articles.
-    You can save information in Vector store for later use.
-    When saving articles in Vector store, you store every sentence as a separate document and 
-    you only save first 10 sentences.
-  `,
-  tools: {
-    lookupWikipedia,
-    saveDocumentInVectorStore,
-  },
-})
-
-const reportCompiler = agent({
-  description: `
-    You are skilled at compiling information from various sources into a coherent report.
-    You can search for specific sentences in Vector database.
-  `,
-  tools: {
-    searchInVectorStore,
-  },
-})
-
-const wikipediaResearch = workflow({
-  team: { wikipediaIndexer, reportCompiler },
-  description: `
-    Find information about John III Sobieski on Wikipedia and save it in Vector store.
-    Lookup sentences related to the following topics:
-     - "Dates of reign as King of Poland"
-     - "John III education"
-  `,
-  knowledge: `
-    Each document in Vector store is a sentence from the Wikipedia article.
-  `,
-  output: `
-    List of sentences looked up for each topic. Each sentence should be in separate bullet point.
-  `,
-  snapshot: logger,
-})
+import { wikipediaResearch } from './wikipedia_vector.config.js'
 
 const result = await teamwork(wikipediaResearch)
 
diff --git a/packages/bdd/README.md b/packages/bdd/README.md
new file mode 100644
index 0000000..e257b29
--- /dev/null
+++ b/packages/bdd/README.md
@@ -0,0 +1,261 @@
+# BDD Testing with Fabrice AI
+
+This guide provides an example of how to write BDD (Behavior-Driven Development) tests using the Fabrice AI framework. The example is based on the `library_photo_to_website.test.ts` file.
+
+## Example: Library Photo to Website
+
+This example demonstrates how to test a workflow that converts a photo of a library into a browsable web catalog. [See full example](../../example/src/library_photo_to_website.test.ts).
+
+### Step-by-Step Guide
+
+1. **Import necessary modules and dependencies:**
+
+```typescript
+import 'dotenv/config'
+
+import { suite, test } from '@fabrice-ai/bdd/suite'
+import { testwork } from '@fabrice-ai/bdd/testwork'
+import fs from 'fs/promises'
+
+import { bookLibraryWorkflow, outputPath, workingDir } from './library_photo_to_website.config.js'
+```
+
+This example somewhat defines the rule convention of saving the workflow in the `*.config.ts` files - so it will be reusable - between tests and executable code. 
+
+Full set of executable/test/workflow files is:
+1. `example/src/library_photo_to_website.config.ts` - workflow definition,
+2. `example/src/library_photo_to_website.test.ts` - test suite,
+3. `example/src/library_photo_to_website.ts` - executable code.
+
+Having this in mind one could use the following commands to run:
+
+- Running tests:
+```ts
+$ tsx library_photo_to_website.test.ts
+```
+
+- Running workflow:
+```ts
+$ tsx library_photo_to_website.ts
+```
+
+
+2. **Define the test suite and test cases:**
+
+```ts
+const testResults = await testwork(
+  bookLibraryWorkflow,
+  suite({
+    description: 'Black box testing suite',
+    team: {
+      librarian: [
+        test(
+          '1_vision',
+          'Librarian should use the vision tool to OCR the photo of the book library to text'
+        ),
+      ],
+      webmaster: [
+        test(
+          '2_listFilesFromDirectory',
+          'Webmaster should list the files from working directory using "listFilesFromDirectory" tool'
+        ),
+        test(
+          '3_saveFile',
+          `Webmaster should modify and save final HTML to ${outputPath} file using "saveFile" tool`
+        ),
+      ],
+    },
+    workflow: [
+      test(
+        '4_search_template',
+        `Webmaster should search and MUST choose the "book_library_template.html" template from inside the ${workingDir} directory.`
+      ),
+      test(
+        '5_finalOutput',
+        'Final list of the books should be at least 5 books long and saved to the HTML file'
+      ),
+      test(
+        '6_finalOutput',
+        `Final output consist "Female Masculinity" title in the ${outputPath} file`,
+        async (workflow, state) => {
+          const htmlContent = await fs.readFile(outputPath, 'utf-8')
+          return {
+            reasoning: "Output file includes the 'Female Masculinity' title",
+            passed: htmlContent.includes('Female Masculinity'),
+            id: '6_finalOutput',
+          }
+        }
+      ),
+    ],
+  })
+)
+
+```
+
+3. **Handle the results:**
+
+```ts
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
+```
+
+## Running the Tests
+
+To run the tests, execute the following command:
+
+```ts
+$ tsx library_photo_to_website.test.ts
+```
+
+This will run the test suite and output the results to the console.
+
+## API
+
+The testing framework API is pretty straightforward.
+
+### `testwork`
+
+Runs the given workflow and continues iterating over the workflow until it finishes. If you handle running tools manually, you can set `runTools` to false.
+
+#### Parameters
+
+- `workflow: Workflow`: The workflow to be tested.
+- `suite: TestSuite`: The test suite containing the test cases.
+- `state: WorkflowState`: The initial state of the workflow. Defaults to `rootState(workflow)`.
+- `runTools: boolean`: Whether to run tools automatically. Defaults to `true`.
+
+#### Returns
+
+- `Promise<TestSuiteResult>`: The overall result of the test suite.
+
+#### Example Usage
+
+```ts
+import { testwork } from '@fabrice-ai/bdd/testwork'
+const testResults = await testwork(
+  bookLibraryWorkflow,
+  suite({ ... })
+)
+if (!testResults.passed) {
+  console.log('🚨 Test suite failed')
+  process.exit(-1)
+} else {
+  console.log('✅ Test suite passed')
+  process.exit(0)
+}
+```
+
+### `suite`
+
+Creates a test suite with the given options.
+
+#### Parameters
+
+- `options: TestSuiteOptions`: The options for creating the test suite.
+
+#### Returns
+
+- `TestSuite`: The created test suite.
+
+#### Example Usage
+
+```ts
+import { suite, test } from '@fabrice-ai/bdd/suite'
+
+const myTestSuite = suite({
+  description: 'Example test suite',
+  workflow: [
+    test('1_exampleTest', 'This is an example test case'),
+  ],
+  team: {
+    exampleAgent: [
+      test('2_exampleAgentTest', 'This is an example test case for an agent'),
+    ],
+  },
+})
+```
+
+
+### `test`
+Creates a test case with the given id, description, and optional run function.
+
+#### Parameters
+`id: string`: The unique identifier for the test case.
+`testCase: string`: The description of the test case.
+r`un?: ((workflow: Workflow, state: WorkflowState) => Promise<SingleTestResult>) | null`: The optional function to run the test case.
+
+### Returns
+`TestCase`: The created test case.
+
+### Example usage
+
+```ts
+import { test } from '@fabrice-ai/bdd/suite'
+
+const exampleTestCase = test('1_exampleTest', 'This is an example test case')
+
+const exampleAgentTestCase = test(
+  '2_exampleAgentTest',
+  'This is an example test case for an agent',
+  async (workflow, state) => {
+    // Custom test logic
+    return {
+      passed: true,
+      reasoning: 'Test passed successfully',
+      id: '2_exampleAgentTest',
+    }
+  }
+)
+```
+
+## Mocking tools
+
+You are able to very easily mock-up the tools used by the agents. For example: tools requesting user attention, or answers could be mocked using the LLM as answering machines - to keep the tests automatic.
+
+Here is just a quick example from the [medical_survey.test.ts](../../example/src/medical_survey.test.ts):
+
+```ts
+export const askUserMock = tool({
+  description: 'Tool for asking user a question',
+  parameters: z.object({
+    query: z.string().describe('The question to ask the user'),
+  }),
+  execute: async ({ query }, { provider }): Promise<string> => {
+    const response = await provider.chat({
+      messages: [
+        {
+          role: 'system',
+          content: `We are role playing - a nurse is asking a patient about their symptoms
+          and the patient is answering. The nurse will ask you a question and you should answer it.
+          Figure out something realistic! It's just a play!`,
+        },
+        {
+          role: 'user',
+          content: 'Try to answer this question in a single line: ' + query,
+        },
+      ],
+      response_format: {
+        result: z.object({
+          answer: z.string().describe('Answer to the question'),
+        }),
+      },
+    })
+    console.log(`😳 Mocked response: ${response.value.answer}\n`)
+    return Promise.resolve(response.value.answer)
+  },
+})
+
+preVisitNoteWorkflow.team['nurse'].tools = {
+  askPatient: askUserMock,
+}
+```
+
+
+## Conclusion
+
+This example demonstrates how to write BDD tests using the Fabrice AI framework. By defining a test suite and test cases, you can validate the behavior of your workflows and ensure they meet the expected requirements. ```
\ No newline at end of file
diff --git a/packages/bdd/package.json b/packages/bdd/package.json
new file mode 100644
index 0000000..d65bd8f
--- /dev/null
+++ b/packages/bdd/package.json
@@ -0,0 +1,49 @@
+{
+  "name": "@fabrice-ai/bdd",
+  "description": "Tools that help you test AI agents",
+  "author": "Piotr Karwatka <piotr@catchthetornado.com>",
+  "scripts": {
+    "prepare": "bun run build",
+    "clean": "rm -rf dist",
+    "build": "tsup-node"
+  },
+  "exports": {
+    "./*": {
+      "bun": "./src/*.ts",
+      "types": "./dist/types/*.d.ts",
+      "require": "./dist/*.cjs",
+      "import": "./dist/*.js"
+    }
+  },
+  "type": "module",
+  "dependencies": {
+    "@clack/prompts": "^0.8.2",
+    "axios": "^1.7.9",
+    "chalk": "^5.3.0",
+    "fabrice-ai": "0.5.2",
+    "zod": "^3.23.8"
+  },
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/callstackincubator/ai-agent-framework.git",
+    "directory": "packages/tools"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "keywords": [
+    "ai",
+    "ai agents",
+    "ai tools"
+  ],
+  "files": [
+    "dist",
+    "src",
+    "README.md"
+  ],
+  "engines": {
+    "node": ">=22"
+  },
+  "version": "0.5.2"
+}
diff --git a/packages/bdd/src/suite.ts b/packages/bdd/src/suite.ts
new file mode 100644
index 0000000..5f83aeb
--- /dev/null
+++ b/packages/bdd/src/suite.ts
@@ -0,0 +1,109 @@
+import { WorkflowState } from 'fabrice-ai/state'
+import { Workflow } from 'fabrice-ai/workflow'
+
+/**
+ * Represents a single test case.
+ */
+export type TestCase = {
+  case: string
+  id: string
+  run: ((workflow: Workflow, state: WorkflowState) => Promise<SingleTestResult>) | null
+}
+
+/**
+ * Represents a test suite containing multiple test cases.
+ */
+export type TestSuite = {
+  description: string
+  workflow: TestCase[]
+  team: {
+    [key: string]: TestCase[]
+  }
+}
+
+/**
+ * Options for creating a test suite.
+ */
+export type TestSuiteOptions = TestSuite
+
+/**
+ * Represents the result of a single test case.
+ */
+export type SingleTestResult = {
+  passed: boolean
+  reasoning: string
+  id: string
+}
+
+/**
+ * Represents the success result of a test suite.
+ */
+export type TestResultsSuccess = {
+  tests: SingleTestResult[]
+}
+
+/**
+ * Represents the failure result of a test suite.
+ */
+export type TestResultsFailure = { reasoning: string; id: string }
+
+/**
+ * Represents the results of a test suite.
+ */
+export type TestResults = TestResultsSuccess | TestResultsFailure
+
+/**
+ * Represents the overall result of a test suite.
+ */
+export type TestSuiteResult = {
+  passed: boolean
+  results: TestResults[]
+}
+
+const defaults = {
+  passed: false,
+}
+
+/**
+ * Represents a request to run a test suite.
+ */
+export type TestRequest = {
+  workflow: Workflow
+  state: WorkflowState
+  teamRouting: Array<string>
+  requestedFor?: string
+  tests: TestCase[]
+}
+
+/**
+ * Creates a test suite with the given options.
+ *
+ * @param options - The options for creating the test suite.
+ * @returns The created test suite.
+ */
+export const suite = (options: TestSuiteOptions): TestSuite => {
+  return {
+    ...defaults,
+    ...options,
+  }
+}
+
+/**
+ * Creates a test case with the given id, description, and optional run function.
+ *
+ * @param id - The unique identifier for the test case.
+ * @param testCase - The description of the test case.
+ * @param run - The optional function to run the test case.
+ * @returns The created test case.
+ */
+export const test = (
+  id: string,
+  testCase: string,
+  run?: ((workflow: Workflow, state: WorkflowState) => Promise<SingleTestResult>) | null
+): TestCase => {
+  return {
+    id,
+    case: testCase,
+    run: run || null,
+  }
+}
diff --git a/packages/bdd/src/testwork.ts b/packages/bdd/src/testwork.ts
new file mode 100644
index 0000000..990b091
--- /dev/null
+++ b/packages/bdd/src/testwork.ts
@@ -0,0 +1,210 @@
+import chalk from 'chalk'
+import s from 'dedent'
+import { iterate } from 'fabrice-ai/iterate'
+import { assistant, system, user } from 'fabrice-ai/messages'
+import { rootState, WorkflowState } from 'fabrice-ai/state'
+import { teamwork } from 'fabrice-ai/teamwork'
+import { logger, Telemetry } from 'fabrice-ai/telemetry'
+import { isCoreTeam, Workflow } from 'fabrice-ai/workflow'
+import { z } from 'zod'
+
+import { SingleTestResult, TestRequest, TestResults, TestSuite, TestSuiteResult } from './suite.js'
+
+const makeTestingVisitor = (
+  workflow: Workflow,
+  suite: TestSuite
+): {
+  testingVisitor: Telemetry
+  testRequests: TestRequest[]
+} => {
+  const testRequests: TestRequest[] = []
+  const teamRouting = new Array<string>()
+  const testingVisitor: Telemetry = async ({ prevState, nextState }) => {
+    if (prevState === nextState) return
+
+    if (!isCoreTeam(nextState.agent)) teamRouting.push(nextState.agent)
+
+    if (
+      nextState.status === 'finished' &&
+      (nextState.agent === 'supervisor' || nextState.agent === 'finalBoss')
+    ) {
+      // test entire workflow
+      testRequests.push({ workflow, state: nextState, tests: suite.workflow, teamRouting })
+    }
+
+    if (nextState.status === 'finished' && suite.team[nextState.agent]) {
+      // test single agent - prevState is internal agent state
+      console.log(`🧪 Requesting test suite for agent [${nextState.agent}]\n`)
+      testRequests.push({
+        workflow,
+        state: prevState,
+        tests: suite.team[nextState.agent],
+        requestedFor: nextState.agent,
+        teamRouting: [],
+      }) // add it only once
+    }
+    // printTree(nextState)
+    return logger({ prevState, nextState })
+  }
+  return { testingVisitor, testRequests }
+}
+
+export async function validate(req: TestRequest): Promise<TestResults> {
+  // evaluate test cases every iterate call - however it could be potentially optimized
+  // to run once at the end.
+  const { workflow, state, tests, teamRouting } = req
+
+  const testRequest = [
+    system(s`
+    You are a LLM test agent.
+
+    Your job is to go thru test cases and evaluate them against the current state.
+    If test case is satisfied mark it passed.
+
+    If you cannot mark the test case as passed, please return it as a unpassed by default.
+
+    Here is the test suite:
+
+    <suite>
+      ${tests
+        .filter((test) => test.run === null) // only run tests that are not defined
+        .map((test) => {
+          return `<test>
+                      <id>${test.id}</id>
+                      <case>${test.case}</case>                     
+                  </test>`
+        })}
+    </suite>
+  `),
+    assistant('What have been done so far?'),
+    user(`Here is the work flow so far:`),
+    ...state.messages,
+    assistant('What was the agent routing?'),
+    user(teamRouting.join(' => ')),
+    assistant(`Is there anything else I need to know?`),
+    workflow.knowledge
+      ? user(`Here is all the knowledge available: ${workflow.knowledge}`)
+      : user(`No, I do not have any additional information.`),
+  ]
+  const suiteResults = await workflow.provider.chat({
+    messages: testRequest,
+    response_format: {
+      suite: z.object({
+        tests: z.array(
+          z.object({
+            id: z.string().describe('The id of the test case'),
+            reasoning: z.string().describe('The reason - why this test passed or not'),
+            passed: z.boolean().describe('The test case is passed or not'),
+          })
+        ),
+      }),
+      error: z.object({
+        id: z.string().describe('The id of the test case'),
+        reasoning: z.string().describe('The reason why you cannot complete the tests'),
+      }),
+    },
+  })
+
+  const testRunners = tests
+    .filter((test) => test.run !== null)
+    .map((test) => {
+      // @ts-ignore
+      return test.run(workflow, state)
+    })
+
+  const subResults = await Promise.all(testRunners)
+
+  if ('tests' in suiteResults.value) {
+    return {
+      tests: [...suiteResults.value.tests, ...subResults],
+    }
+  }
+
+  return suiteResults.value // error - no test results, just the `reasoning` for why it failed
+}
+
+const printTestResult = (
+  level: number,
+  testId: string,
+  icon: string,
+  message: string,
+  reason: string
+) => {
+  const indent = '  '.repeat(level)
+  const arrow = level > 0 ? '└─▶ ' : ''
+  console.log(`${indent}${arrow}${icon}${chalk.bold(testId)}: ${message}`)
+  console.log(`${indent} 🧠 ${chalk.dim(reason)}`)
+}
+
+export const displayTestResults = (results: SingleTestResult[]) => {
+  console.log('🧪 Test results: ')
+  results.map((testResult) => {
+    printTestResult(
+      2,
+      testResult.id,
+      `${testResult.passed ? '✅' : '🚨'}`,
+      `${testResult.passed ? 'PASSED' : 'FAIL'}`,
+      testResult.reasoning
+    )
+  })
+}
+/**
+ * Teamwork runs given workflow and continues iterating over the workflow until it finishes.
+ * If you handle running tools manually, you can set runTools to false.
+ */
+export async function testwork(
+  workflow: Workflow,
+  suite: TestSuite,
+  state: WorkflowState = rootState(workflow),
+  runTools: boolean = true
+): Promise<TestSuiteResult> {
+  const { testingVisitor, testRequests } = makeTestingVisitor(workflow, suite)
+  workflow.snapshot = testingVisitor
+  const nextState = await teamwork(workflow, await iterate(workflow, state), runTools)
+  if (nextState.status === 'finished') {
+    const overallResults = await Promise.all(
+      testRequests.map((testRequest) => {
+        console.log(`🧪 Running test suite [${testRequest.tests.map((t) => t.id).join(', ')}]\n`)
+        return validate(testRequest)
+      })
+    )
+
+    const finalResults = overallResults.flatMap((result) => {
+      if ('tests' in result) return result.tests
+      else return [{ ...result, passed: false }] // case of general issue with the whole set of tests
+    })
+    const requiredAgentCalls = Object.keys(suite.team)
+    const missingAgentCalls = requiredAgentCalls
+      .map((requiredAgent) =>
+        !testRequests.find((req) => req.requestedFor === requiredAgent) ? requiredAgent : null
+      )
+      .filter((agent) => agent !== null)
+
+    if (missingAgentCalls.length > 0) {
+      console.log(`🚨 Missing test suites for agents: ${missingAgentCalls}\n`)
+      finalResults.push({
+        passed: false,
+        reasoning: 'Missing test suites for agents: ' + missingAgentCalls.join(', '),
+        id: 'missing_agent_calls',
+      })
+      missingAgentCalls.forEach((agent) => {
+        suite.team[agent].forEach((test) => {
+          finalResults.push({
+            passed: false,
+            reasoning: 'Missing call for agent ' + agent,
+            id: test.id,
+          })
+        })
+      })
+    }
+
+    displayTestResults(finalResults)
+    return { passed: finalResults.every((test) => test.passed), results: overallResults }
+  }
+
+  if (nextState.status === 'failed') {
+    throw Error('Workflow did not finish successfully')
+  }
+
+  return await testwork(workflow, suite, nextState, runTools)
+}
diff --git a/packages/bdd/tsconfig.json b/packages/bdd/tsconfig.json
new file mode 100644
index 0000000..7513033
--- /dev/null
+++ b/packages/bdd/tsconfig.json
@@ -0,0 +1,4 @@
+{
+  "extends": "../../tsconfig.json",
+  "include": ["./src"]
+}
diff --git a/packages/bdd/tsup.config.ts b/packages/bdd/tsup.config.ts
new file mode 100644
index 0000000..3c968b6
--- /dev/null
+++ b/packages/bdd/tsup.config.ts
@@ -0,0 +1,10 @@
+import { defineConfig } from 'tsup'
+
+export default defineConfig({
+  entry: ['./src/*.ts'],
+  format: ['cjs', 'esm'],
+  target: 'node20',
+  splitting: false,
+  clean: true,
+  dts: true,
+})
diff --git a/packages/create-fabrice-ai/src/index.ts b/packages/create-fabrice-ai/src/index.ts
index 3d904ae..c4b172c 100644
--- a/packages/create-fabrice-ai/src/index.ts
+++ b/packages/create-fabrice-ai/src/index.ts
@@ -73,13 +73,22 @@ const template = await select({
   options: [
     {
       value: {
-        files: ['src/surprise_trip.ts', 'src/tools/wikipedia.ts'],
+        files: [
+          'src/surprise_trip.ts',
+          'src/surprise_trip.test.ts',
+          'src/surprise_trip.config.ts',
+          'src/tools/wikipedia.ts',
+        ],
       },
       label: 'Surprise Trip Planner - travel agent creating personalized city adventures',
     },
     {
       value: {
-        files: ['src/medical_survey.ts', 'src/medical_survey/workflow.ts'],
+        files: [
+          'src/medical_survey.ts',
+          'src/medical_survey.test.ts',
+          'src/medical_survey.config.ts',
+        ],
       },
       label: 'Medical Survey - Pre-visit patient questionnaire with report generation',
     },
@@ -87,6 +96,8 @@ const template = await select({
       value: {
         files: [
           'src/library_photo_to_website.ts',
+          'src/library_photo_to_website.test.ts',
+          'src/library_photo_to_website.config.ts',
           'assets/photo-library.jpg',
           'assets/book_library_template.html',
         ],
@@ -95,33 +106,42 @@ const template = await select({
     },
     {
       value: {
-        files: ['src/ecommerce_product_description.ts', 'assets/example-sneakers.jpg'],
+        files: [
+          'src/ecommerce_product_description.ts',
+          'src/ecommerce_product_description.test.ts',
+          'src/ecommerce_product_description.config.ts',
+          'assets/example-sneakers.jpg',
+        ],
       },
       label:
         'E-commerce Product Description - Convert product photos into compelling store listings',
     },
     {
       value: {
-        files: ['src/news_wrap_up.ts'],
+        files: ['src/news_wrap_up.ts', 'src/news_wrap_up.config.ts', 'src/news_wrap_up.test.ts'],
       },
       label: 'News Wrap Up - Weekly news digest generator with smart summaries',
     },
     {
       value: {
-        files: ['src/github_trending.ts'],
-      },
-      label: 'GitHub Trending - Track and summarize hot Python projects on GitHub',
-    },
-    {
-      value: {
-        files: ['src/github_trending_vector.ts', 'src/tools/askUser.ts'],
+        files: [
+          'src/github_trending_vector.ts',
+          'src/github_trending_vector.config.ts',
+          'src/github_trending_vector.test.ts',
+          'src/tools/askUser.ts',
+        ],
       },
       label:
         'GitHub Trending + Vector Store - Track and summarize hot Typescript projects on GitHub + get project details from vector store',
     },
     {
       value: {
-        files: ['src/wikipedia_vector.ts', 'src/tools/wikipedia.ts'],
+        files: [
+          'src/wikipedia_vector.ts',
+          'src/wikipedia_vector.test.ts',
+          'src/wikipedia_vector.config.ts',
+          'src/tools/wikipedia.ts',
+        ],
       },
       label: 'Wikipedia Vector - Search and summarize Wikipedia articles',
     },
diff --git a/packages/framework/src/agent.ts b/packages/framework/src/agent.ts
index 20094c3..7b8b4bf 100644
--- a/packages/framework/src/agent.ts
+++ b/packages/framework/src/agent.ts
@@ -1,7 +1,7 @@
 import s from 'dedent'
 import { z } from 'zod'
 
-import { assistant, getSteps, Message, system, toolCalls, user } from './messages.js'
+import { assistant, Conversation, getSteps, Message, system, toolCalls, user } from './messages.js'
 import { Provider } from './models.js'
 import { finish, WorkflowState } from './state.js'
 import { Tool } from './tool.js'
@@ -102,7 +102,18 @@ export const agent = (options: AgentOptions = {}): Agent => {
           }
         }
 
-        return finish(state, agentResponse)
+        const prevState: WorkflowState = {
+          ...state,
+          status: 'running',
+          messages: [
+            ...state.messages,
+            agentResponse,
+            user(response.value.next_step),
+          ] as Conversation,
+        }
+        const nextState = finish(state, agentResponse)
+        workflow.snapshot({ prevState, nextState })
+        return nextState
       }),
   }
 }
diff --git a/packages/framework/src/telemetry.ts b/packages/framework/src/telemetry.ts
index 7393e9f..0b95c78 100644
--- a/packages/framework/src/telemetry.ts
+++ b/packages/framework/src/telemetry.ts
@@ -15,7 +15,7 @@ export const logger: Telemetry = ({ prevState, nextState }) => {
   if (prevState === nextState) return
 
   const getStatusText = (state: WorkflowState) => {
-    if (state.agent === 'supervisor') {
+    if (state.agent === 'supervisor' && (state.status === 'idle' || state.status === 'running')) {
       return 'Looking for next task...'
     }
     if (state.agent === 'resourcePlanner') {