minerva.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Minerva: AI Guardian for Scam Protection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook implements a multi-agent system for analyzing images (screenshots) to identify scam attempts, and provide personalized scam prevention. It uses [AutoGen](https://github.com/microsoft/autogen/) to orchestrate various specialized agents that work together.\n",
    "\n",
    "Benefits:\n",
    "- Automates the process of identifying suspicious scam patterns.\n",
    "- Prevent Financial Loss\n",
    "- Save Time: Early scam detection reduces the number of claims filed by end-users."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Install Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "%pip install -q pyautogen pillow pytesseract"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "import autogen\n",
    "from IPython.display import Image as IPImage\n",
    "from IPython.display import display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from dotenv import load_dotenv, find_dotenv\n",
    "\n",
    "load_dotenv(find_dotenv())\n",
    "\n",
    "llm_config = [\n",
    "    {\n",
    "        \"model\": \"gpt-4o-mini\",\n",
    "        \"api_key\": os.getenv(\"OPENAI_API_KEY\")\n",
    "    }\n",
    "]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tools Definition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ocr(image_path: str) -> str:\n",
    "    from PIL import Image\n",
    "    import pytesseract\n",
    "\n",
    "    try:\n",
    "        image = Image.open(image_path)\n",
    "        text = pytesseract.image_to_string(image)\n",
    "        return text\n",
    "    except Exception as e:\n",
    "        return f\"Error in text extraction: {str(e)}\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Agents Creation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_agents():\n",
    "    \"\"\"Create and initialize the specialized agents.\"\"\"\n",
    "    \n",
    "    # OCR Agent - Extracts and processes text from images\n",
    "    ocr_agent = autogen.AssistantAgent(\n",
    "        name=\"OCR_Specialist\",\n",
    "        system_message=\"\"\"You are an OCR specialist. Your role is to:\n",
    "        1. Extract text from an image path using pytesseract\n",
    "        2. Clean and format the extracted text for further analysis\"\"\",\n",
    "        llm_config={\"config_list\": llm_config}\n",
    "    )\n",
    "\n",
    "    # Content Analysis Agent - Evaluates text content for scam signals\n",
    "    content_agent = autogen.AssistantAgent(\n",
    "        name=\"Content_Analyst\",\n",
    "        system_message=\"\"\"You are a content analysis specialist. Your role is to:\n",
    "        1. Analyze text for common scam patterns\n",
    "        2. Identify and analyze URLs, phone numbers, or other contact information\n",
    "        3. Identify urgency indicators, threats, or pressure tactics\n",
    "        5. Check for inconsistencies in messaging\n",
    "        6. Evaluate legitimacy of any claims or offers\"\"\",\n",
    "        llm_config={\"config_list\": llm_config}\n",
    "    )\n",
    "\n",
    "    # Decision Making Agent - Makes final determination based on all analyses\n",
    "    decision_agent = autogen.AssistantAgent(\n",
    "        name=\"Decision_Maker\",\n",
    "        system_message=\"\"\"You are the final decision maker. Your role is to:\n",
    "        1. Coordinate with other agents to gather all necessary information\n",
    "        2. Make a final determination on scam probability\n",
    "        3. Provide detailed explanation of the decision\n",
    "        4. End your explanation with the label as 'TASK_COMPLETE' when done\"\"\",\n",
    "        llm_config={\"config_list\": llm_config}\n",
    "    )\n",
    "\n",
    "    summary_agent = autogen.AssistantAgent(\n",
    "        name=\"Summary_Agent\",\n",
    "        system_message=\"\"\"You are a communication specialist who creates clear, concise summaries of technical analyses. Your role is to:\n",
    "        1. Synthesize the findings of a scam assessment into user-friendly language\n",
    "        2. Highlight the most important points that users need to know\n",
    "        3. Provide actionable recommendations\n",
    "        4. Use clear, non-technical language while maintaining accuracy\n",
    "        5. Format information in a way that's easy to read and understand\"\"\",\n",
    "        llm_config={\"config_list\": llm_config}\n",
    "    )\n",
    "\n",
    "    # User proxy for automated interaction and code execution\n",
    "    user_proxy = autogen.UserProxyAgent(\n",
    "        name=\"User_Proxy\",\n",
    "        human_input_mode=\"NEVER\",\n",
    "        max_consecutive_auto_reply=5,\n",
    "        is_termination_msg=lambda x: \"TASK_COMPLETE\" in x.get(\"content\", \"\"),\n",
    "    )\n",
    "\n",
    "    ocr_agent.register_for_llm(name=\"ocr\", description=\"Extracts text from an image path\")(ocr)\n",
    "    user_proxy.register_for_execution(name=\"ocr\")(ocr)\n",
    "    \n",
    "    return ocr_agent, content_agent, decision_agent, user_proxy "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Workflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ScamDetectionWorkflow:\n",
    "    def __init__(self):\n",
    "        self.ocr_agent, self.content_agent, self.decision_agent, self.user_proxy = create_agents()\n",
    "        \n",
    "    def analyze(self, image_path):\n",
    "        \"\"\"Coordinate the multi-agent analysis.\n",
    "        \"\"\"\n",
    "        image_path = \"./samples/02.giftcard.message.scam.png\"\n",
    "        \n",
    "        groupchat = autogen.GroupChat(\n",
    "            agents=[self.ocr_agent, self.content_agent, self.decision_agent, self.user_proxy],\n",
    "            messages=[],\n",
    "            max_round=10\n",
    "        )\n",
    "        manager = autogen.GroupChatManager(groupchat=groupchat)\n",
    "\n",
    "        # Start the collaborative analysis\n",
    "        messages = self.user_proxy.initiate_chat(\n",
    "            manager,\n",
    "            message=f\"\"\"Please analyze the content of an image for scam indicators:\n",
    "            1. OCR Agent: Extract text from this image: {image_path}\n",
    "            2. Content Agent: Evaluate the messaging and claims\n",
    "            3. Decision Maker: Synthesize all analyses and make final determination\"\"\",\n",
    "        )\n",
    "\n",
    "        return messages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[33mUser_Proxy\u001b[0m (to chat_manager):\n",
      "\n",
      "Please analyze the content of an image for scam indicators:\n",
      "            1. OCR Agent: Extract text from this image: ./samples/02.giftcard.message.scam.png\n",
      "            2. Content Agent: Evaluate the messaging and claims\n",
      "            3. Decision Maker: Synthesize all analyses and make final determination\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: OCR_Specialist\n",
      "\u001b[0m\n",
      "\u001b[33mOCR_Specialist\u001b[0m (to chat_manager):\n",
      "\n",
      "\u001b[32m***** Suggested tool call (call_2CqBTFMP05FHbF5nQg93Kn5U): ocr *****\u001b[0m\n",
      "Arguments: \n",
      "{\"image_path\":\"./samples/02.giftcard.message.scam.png\"}\n",
      "\u001b[32m********************************************************************\u001b[0m\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: User_Proxy\n",
      "\u001b[0m\n",
      "\u001b[35m\n",
      ">>>>>>>> EXECUTING FUNCTION ocr...\u001b[0m\n",
      "\u001b[33mUser_Proxy\u001b[0m (to chat_manager):\n",
      "\n",
      "\u001b[32m***** Response from calling tool (call_2CqBTFMP05FHbF5nQg93Kn5U) *****\u001b[0m\n",
      "Congratulations!\n",
      "You've won a $1,000\n",
      "Walmart gift card. Go\n",
      "\n",
      "http://bit.ly/123456\n",
      "tp claim now.\n",
      "\f\n",
      "\u001b[32m**********************************************************************\u001b[0m\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: OCR_Specialist\n",
      "\u001b[0m\n",
      "\u001b[33mOCR_Specialist\u001b[0m (to chat_manager):\n",
      "\n",
      "\u001b[32m***** Suggested tool call (call_9KDRpqdLLwaU4kwtmsCPfXLN): ocr *****\u001b[0m\n",
      "Arguments: \n",
      "{\"image_path\": \"./samples/02.giftcard.message.scam.png\"}\n",
      "\u001b[32m********************************************************************\u001b[0m\n",
      "\u001b[32m***** Suggested tool call (call_4P6ZQAv9jirUcmieUUSavrIX): ocr *****\u001b[0m\n",
      "Arguments: \n",
      "{\"image_path\": \"./samples/02.giftcard.message.scam.png\"}\n",
      "\u001b[32m********************************************************************\u001b[0m\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: User_Proxy\n",
      "\u001b[0m\n",
      "\u001b[35m\n",
      ">>>>>>>> EXECUTING FUNCTION ocr...\u001b[0m\n",
      "\u001b[35m\n",
      ">>>>>>>> EXECUTING FUNCTION ocr...\u001b[0m\n",
      "\u001b[33mUser_Proxy\u001b[0m (to chat_manager):\n",
      "\n",
      "\u001b[32m***** Response from calling tool (call_9KDRpqdLLwaU4kwtmsCPfXLN) *****\u001b[0m\n",
      "Congratulations!\n",
      "You've won a $1,000\n",
      "Walmart gift card. Go\n",
      "\n",
      "http://bit.ly/123456\n",
      "tp claim now.\n",
      "\f\n",
      "\u001b[32m**********************************************************************\u001b[0m\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m***** Response from calling tool (call_4P6ZQAv9jirUcmieUUSavrIX) *****\u001b[0m\n",
      "Congratulations!\n",
      "You've won a $1,000\n",
      "Walmart gift card. Go\n",
      "\n",
      "http://bit.ly/123456\n",
      "tp claim now.\n",
      "\f\n",
      "\u001b[32m**********************************************************************\u001b[0m\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: OCR_Specialist\n",
      "\u001b[0m\n",
      "\u001b[33mOCR_Specialist\u001b[0m (to chat_manager):\n",
      "\n",
      "### Extracted Text\n",
      "The extracted text from the image is:\n",
      "\n",
      "```\n",
      "Congratulations!\n",
      "You've won a $1,000\n",
      "Walmart gift card. Go\n",
      "http://bit.ly/123456\n",
      "tp claim now.\n",
      "```\n",
      "\n",
      "### Analysis of Content\n",
      "1. **Claims of Winning**: The message claims that the recipient has won a significant prize ($1,000 Walmart gift card). Such claims often come without entry in contests, which is a common indicator of scams.\n",
      "\n",
      "2. **Urgency and Action Directive**: The phrase \"Go tp claim now\" creates urgency, pushing the reader to act quickly. Scammers often employ urgency to provoke emotional responses leading to hasty decisions.\n",
      "\n",
      "3. **Use of Link**: The URL uses a shortened link (bit.ly). Scammers frequently use such links to obscure the final destination, making it hard for users to see where they are being directed. This is a major red flag.\n",
      "\n",
      "4. **Lack of Personalization**: The message does not address the recipient by name or provide any details about how they \"won,\" which is typical of generic scam messages.\n",
      "\n",
      "5. **Formatting and Miscues**: The text contains formatting issues, such as \"tp\" instead of \"to,\" which is unprofessional and suggests a lack of authenticity.\n",
      "\n",
      "### Final Determination\n",
      "Based on the characteristics of the extracted message:\n",
      "\n",
      "- The claim of winning a prize without any entry process.\n",
      "- The use of urgency and action-oriented language.\n",
      "- The inclusion of a suspicious shortened URL.\n",
      "- The generic nature of the communication.\n",
      "- The presence of grammatical errors and poor formatting.\n",
      "\n",
      "**Conclusion**: This message exhibits strong indicators of a scam. It is advisable to disregard the message and not click on any links provided.\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: Content_Analyst\n",
      "\u001b[0m\n",
      "\u001b[33mContent_Analyst\u001b[0m (to chat_manager):\n",
      "\n",
      "### Extracted Text\n",
      "The extracted text from the image is:\n",
      "\n",
      "```\n",
      "Congratulations!\n",
      "You've won a $1,000\n",
      "Walmart gift card. Go\n",
      "http://bit.ly/123456\n",
      "tp claim now.\n",
      "```\n",
      "\n",
      "### Analysis of Content\n",
      "1. **Claims of Winning**: The message claims that the recipient has won a significant prize ($1,000 Walmart gift card). Such claims are commonly associated with scams, especially when there is no previous participation in a contest.\n",
      "\n",
      "2. **Urgency and Action Directive**: The phrase \"Go tp claim now\" suggests an urgent call to action, which is a tactic often used by scammers to provoke quick responses from victims.\n",
      "\n",
      "3. **Use of Link**: The URL utilizes a shortened link (bit.ly), which is often a red flag. Scammers typically use such links to mask the actual web address, making it difficult for users to determine the legitimacy of the site they are being directed to.\n",
      "\n",
      "4. **Lack of Personalization**: The absence of personalization—such as addressing the recipient by name or providing specific context around the \"win\"—is common in scam communications, indicating it is a mass-produced message.\n",
      "\n",
      "5. **Formatting and Miscues**: The phrase \"tp\" instead of \"to\" reflects poor grammar and formatting, adding to the impression that the message is unprofessional and possibly part of a scam scheme.\n",
      "\n",
      "### Final Determination\n",
      "Based on the characteristics of the extracted message:\n",
      "\n",
      "- The claim of winning a prize with no evidence of entry.\n",
      "- The use of urgency in the language.\n",
      "- The presence of a suspicious shortened URL.\n",
      "- The generic nature and lack of personalization in the message.\n",
      "- Grammatical errors suggesting a lack of authenticity.\n",
      "\n",
      "**Conclusion**: This message strongly indicates a scam. It is advisable to ignore the message and refrain from clicking any links provided.\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: Decision_Maker\n",
      "\u001b[0m\n",
      "\u001b[33mDecision_Maker\u001b[0m (to chat_manager):\n",
      "\n",
      "### Final Determination on Scam Probability\n",
      "\n",
      "**Synthesis of Analyses**\n",
      "\n",
      "Both the OCR Agent's extraction and Content Analyst's evaluations reveal consistent and concerning indicators characteristic of fraud:\n",
      "\n",
      "1. **Claims of Winning**: Both analyses highlight the prominent claim of winning a $1,000 Walmart gift card without any prior engagement or contest participation. This is a common tactic in scams aiming to entice individuals with offers that sound too good to be true.\n",
      "\n",
      "2. **Urgency and Action Directive**: The urging phrase \"Go tp claim now\" was noted in both evaluations, showcasing an attempt to create urgency. This pressure is a typical strategy employed by scammers to rush potential victims into making impulsive decisions.\n",
      "\n",
      "3. **Suspicious Link Use**: The presence of a shortened URL (http://bit.ly/123456) serves as a significant red flag in both analyses. Scammers often obscure their true web addresses, using shortened links to direct individuals to potentially harmful sites. \n",
      "\n",
      "4. **Lack of Personalization**: The absence of specific details that pertain to the recipient aligns with conventional scam communication, indicating it is likely mass-produced.\n",
      "\n",
      "5. **Grammatical Issues and Lack of Professionalism**: Errors in language, particularly the misspelling \"tp\" instead of \"to,\" were flagged in both analyses, further emphasizing the unprofessional and potentially fraudulent origin of the message.\n",
      "\n",
      "### Conclusion\n",
      "Based on the strong indicators derived from both content analysis and extraction, it is appropriate to conclude that this message is highly indicative of a scam. Individuals should be cautious, avoid clicking any hyperlinks, and report the message if received via electronic means.\n",
      "\n",
      "**TASK_COMPLETE**\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "\u001b[32m\n",
      "Next speaker: User_Proxy\n",
      "\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "image_path = \"./samples/02.giftcard.message.scam.png\"\n",
    "display(IPImage(filename=image_path))\n",
    "\n",
    "workflow = ScamDetectionWorkflow()\n",
    "results = workflow.analyze(image_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'content': '### Final Determination on Scam Probability\\n'\n",
      "            '\\n'\n",
      "            '**Synthesis of Analyses**\\n'\n",
      "            '\\n'\n",
      "            \"Both the OCR Agent's extraction and Content Analyst's evaluations \"\n",
      "            'reveal consistent and concerning indicators characteristic of '\n",
      "            'fraud:\\n'\n",
      "            '\\n'\n",
      "            '1. **Claims of Winning**: Both analyses highlight the prominent '\n",
      "            'claim of winning a $1,000 Walmart gift card without any prior '\n",
      "            'engagement or contest participation. This is a common tactic in '\n",
      "            'scams aiming to entice individuals with offers that sound too '\n",
      "            'good to be true.\\n'\n",
      "            '\\n'\n",
      "            '2. **Urgency and Action Directive**: The urging phrase \"Go tp '\n",
      "            'claim now\" was noted in both evaluations, showcasing an attempt '\n",
      "            'to create urgency. This pressure is a typical strategy employed '\n",
      "            'by scammers to rush potential victims into making impulsive '\n",
      "            'decisions.\\n'\n",
      "            '\\n'\n",
      "            '3. **Suspicious Link Use**: The presence of a shortened URL '\n",
      "            '(http://bit.ly/123456) serves as a significant red flag in both '\n",
      "            'analyses. Scammers often obscure their true web addresses, using '\n",
      "            'shortened links to direct individuals to potentially harmful '\n",
      "            'sites. \\n'\n",
      "            '\\n'\n",
      "            '4. **Lack of Personalization**: The absence of specific details '\n",
      "            'that pertain to the recipient aligns with conventional scam '\n",
      "            'communication, indicating it is likely mass-produced.\\n'\n",
      "            '\\n'\n",
      "            '5. **Grammatical Issues and Lack of Professionalism**: Errors in '\n",
      "            'language, particularly the misspelling \"tp\" instead of \"to,\" were '\n",
      "            'flagged in both analyses, further emphasizing the unprofessional '\n",
      "            'and potentially fraudulent origin of the message.\\n'\n",
      "            '\\n'\n",
      "            '### Conclusion\\n'\n",
      "            'Based on the strong indicators derived from both content analysis '\n",
      "            'and extraction, it is appropriate to conclude that this message '\n",
      "            'is highly indicative of a scam. Individuals should be cautious, '\n",
      "            'avoid clicking any hyperlinks, and report the message if received '\n",
      "            'via electronic means.\\n'\n",
      "            '\\n'\n",
      "            '**TASK_COMPLETE**',\n",
      " 'name': 'Decision_Maker',\n",
      " 'role': 'user'}\n"
     ]
    }
   ],
   "source": [
    "import pprint\n",
    "\n",
    "pprint.pprint(results.chat_history[-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "\n",
    "\n",
    "with open('results.json', 'w') as json_file:\n",
    "    json.dump(results.__dict__, json_file, indent=4)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}