Spaces:
Running
Running
File size: 29,110 Bytes
4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 4b93b4a ecec503 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 |
{
"cells": [
{
"cell_type": "markdown",
"id": "a684408b",
"metadata": {},
"source": [
"# 04 - Agent Intelligence\n",
"\n",
"## CyberForge AI - Agentic AI Capabilities\n",
"\n",
"This notebook implements intelligent agent decision-making for:\n",
"- Task prioritization and execution planning\n",
"- Confidence-based decision scoring\n",
"- Gemini API integration for reasoning\n",
"- Autonomous threat response workflows\n",
"\n",
"### Alignment:\n",
"- Integrates with desktop app agentic system\n",
"- Supports backend task queue management\n",
"- Provides explainable outputs for user transparency"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "8af0eaa6",
"metadata": {},
"outputs": [],
"source": [
"import json\n",
"import os\n",
"import time\n",
"import numpy as np\n",
"from pathlib import Path\n",
"from typing import Dict, List, Any, Optional, Tuple\n",
"from dataclasses import dataclass, asdict\n",
"from enum import Enum\n",
"import warnings\n",
"warnings.filterwarnings('ignore')\n",
"\n",
"# Load configuration\n",
"config_path = Path(\"notebook_config.json\")\n",
"if not config_path.exists():\n",
" config_path = Path(\"/home/user/app/notebooks/notebook_config.json\")\n",
"with open(config_path) as f:\n",
" CONFIG = json.load(f)\n",
"\n",
"MODELS_DIR = Path(CONFIG[\"datasets_dir\"]).parent / \"models\"\n",
"AGENT_DIR = MODELS_DIR.parent / \"agent\"\n",
"AGENT_DIR.mkdir(exist_ok=True)\n",
"\n",
"print(f\"✓ Configuration loaded\")\n",
"print(f\"✓ Agent output: {AGENT_DIR}\")"
]
},
{
"cell_type": "markdown",
"id": "bbcc7dc9",
"metadata": {},
"source": [
"## 1. Agent Task Definitions"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bc6eb821",
"metadata": {},
"outputs": [],
"source": [
"class TaskPriority(Enum):\n",
" CRITICAL = 4\n",
" HIGH = 3\n",
" MEDIUM = 2\n",
" LOW = 1\n",
" BACKGROUND = 0\n",
"\n",
"@dataclass\n",
"class AgentTask:\n",
" \"\"\"Represents a task the agent can execute\"\"\"\n",
" task_id: str\n",
" task_type: str\n",
" priority: int\n",
" target: str\n",
" context: Dict\n",
" created_at: float\n",
" confidence: float = 0.0\n",
" status: str = \"pending\"\n",
" result: Dict = None\n",
" \n",
" def to_dict(self) -> Dict:\n",
" return asdict(self)\n",
"\n",
"@dataclass\n",
"class AgentDecision:\n",
" \"\"\"Represents an agent decision with reasoning\"\"\"\n",
" action: str\n",
" confidence: float\n",
" reasoning: str\n",
" evidence: List[str]\n",
" risk_level: str\n",
" recommended_follow_up: List[str]\n",
" \n",
" def to_dict(self) -> Dict:\n",
" return asdict(self)\n",
"\n",
"print(\"✓ Task definitions loaded\")"
]
},
{
"cell_type": "markdown",
"id": "c6785168",
"metadata": {},
"source": [
"## 2. Decision Scoring Engine"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "0e0109ae",
"metadata": {},
"outputs": [],
"source": [
"class DecisionScoringEngine:\n",
" \"\"\"\n",
" Calculates confidence scores for agent decisions.\n",
" Combines model predictions with heuristic rules.\n",
" \"\"\"\n",
" \n",
" # Threat severity weights\n",
" SEVERITY_WEIGHTS = {\n",
" 'critical': 1.0,\n",
" 'high': 0.8,\n",
" 'medium': 0.5,\n",
" 'low': 0.3,\n",
" 'info': 0.1\n",
" }\n",
" \n",
" # Evidence type weights\n",
" EVIDENCE_WEIGHTS = {\n",
" 'model_prediction': 0.4,\n",
" 'signature_match': 0.3,\n",
" 'behavioral_pattern': 0.2,\n",
" 'heuristic_rule': 0.1\n",
" }\n",
" \n",
" def __init__(self, confidence_threshold: float = 0.7):\n",
" self.confidence_threshold = confidence_threshold\n",
" \n",
" def calculate_threat_score(self, indicators: List[Dict]) -> Tuple[float, str]:\n",
" \"\"\"Calculate threat score from multiple indicators\"\"\"\n",
" if not indicators:\n",
" return 0.0, 'low'\n",
" \n",
" # Weighted average of indicator scores\n",
" total_weight = 0\n",
" total_score = 0\n",
" \n",
" for indicator in indicators:\n",
" severity = indicator.get('severity', 'low')\n",
" confidence = indicator.get('confidence', 0.5)\n",
" evidence_type = indicator.get('evidence_type', 'heuristic_rule')\n",
" \n",
" weight = self.SEVERITY_WEIGHTS.get(severity, 0.3) * \\\n",
" self.EVIDENCE_WEIGHTS.get(evidence_type, 0.1)\n",
" \n",
" total_weight += weight\n",
" total_score += confidence * weight\n",
" \n",
" if total_weight == 0:\n",
" return 0.0, 'low'\n",
" \n",
" final_score = total_score / total_weight\n",
" \n",
" # Determine risk level\n",
" if final_score >= 0.8:\n",
" risk = 'critical'\n",
" elif final_score >= 0.6:\n",
" risk = 'high'\n",
" elif final_score >= 0.4:\n",
" risk = 'medium'\n",
" elif final_score >= 0.2:\n",
" risk = 'low'\n",
" else:\n",
" risk = 'info'\n",
" \n",
" return final_score, risk\n",
" \n",
" def should_act(self, score: float, task_type: str) -> bool:\n",
" \"\"\"Determine if agent should act based on score\"\"\"\n",
" # Different thresholds for different task types\n",
" thresholds = {\n",
" 'block_threat': 0.85,\n",
" 'quarantine': 0.75,\n",
" 'alert': 0.5,\n",
" 'scan': 0.3,\n",
" 'monitor': 0.1\n",
" }\n",
" \n",
" threshold = thresholds.get(task_type, self.confidence_threshold)\n",
" return score >= threshold\n",
" \n",
" def prioritize_tasks(self, tasks: List[AgentTask]) -> List[AgentTask]:\n",
" \"\"\"Sort tasks by priority and confidence\"\"\"\n",
" return sorted(tasks, \n",
" key=lambda t: (t.priority, t.confidence), \n",
" reverse=True)\n",
"\n",
"scoring_engine = DecisionScoringEngine()\n",
"print(\"✓ Decision Scoring Engine initialized\")"
]
},
{
"cell_type": "markdown",
"id": "572652b4",
"metadata": {},
"source": [
"## 3. Gemini API Integration for Reasoning"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "e68b585e",
"metadata": {},
"outputs": [],
"source": [
"# Gemini Integration - using google-genai (same as ml-services/app/services/gemini_service.py)\n",
"try:\n",
" from google import genai\n",
" GEMINI_AVAILABLE = True\n",
"except ImportError:\n",
" import subprocess\n",
" subprocess.run(['pip', 'install', 'google-genai', '-q'])\n",
" from google import genai\n",
" GEMINI_AVAILABLE = True\n",
"\n",
"class GeminiReasoningEngine:\n",
" \"\"\"\n",
" Uses Gemini API for intelligent threat reasoning.\n",
" Provides explainable AI outputs.\n",
" \"\"\"\n",
" \n",
" SYSTEM_PROMPT = \"\"\"You are a cybersecurity AI agent analyzing security threats.\n",
"Your role is to:\n",
"1. Analyze security indicators and threat patterns\n",
"2. Provide clear, actionable recommendations\n",
"3. Explain your reasoning in a way users can understand\n",
"4. Prioritize user safety while minimizing false positives\n",
"\n",
"Always respond with JSON containing:\n",
"- action: recommended action (block, alert, monitor, allow)\n",
"- confidence: 0.0-1.0 confidence score\n",
"- reasoning: brief explanation\n",
"- evidence: list of supporting evidence\n",
"- risk_level: critical/high/medium/low/info\n",
"- recommended_follow_up: list of next steps\"\"\"\n",
" \n",
" def __init__(self):\n",
" self.api_key = CONFIG.get('gemini_api_key', os.environ.get('GEMINI_API_KEY'))\n",
" self.client = None\n",
" self.model_name = CONFIG.get('gemini_model', os.environ.get('GEMINI_MODEL', 'gemini-2.5-flash'))\n",
" \n",
" if self.api_key:\n",
" try:\n",
" self.client = genai.Client(api_key=self.api_key)\n",
" # Test connection\n",
" test = self.client.models.generate_content(\n",
" model=self.model_name,\n",
" contents=\"Test. Respond with OK.\"\n",
" )\n",
" if test.text:\n",
" print(f\" ✓ Gemini API connected (model: {self.model_name})\")\n",
" else:\n",
" print(\" ⚠ Gemini API responded but with empty text\")\n",
" except Exception as e:\n",
" print(f\" ⚠ Gemini API error: {e}\")\n",
" self.client = None\n",
" else:\n",
" print(\" ⚠ No Gemini API key found (will use fallback reasoning)\")\n",
" \n",
" def analyze_threat(self, threat_data: Dict) -> AgentDecision:\n",
" \"\"\"Analyze threat and generate decision with reasoning\"\"\"\n",
" if self.client:\n",
" return self._gemini_analyze(threat_data)\n",
" else:\n",
" return self._fallback_analyze(threat_data)\n",
" \n",
" def _gemini_analyze(self, threat_data: Dict) -> AgentDecision:\n",
" \"\"\"Use Gemini for threat analysis\"\"\"\n",
" prompt = f\"\"\"{self.SYSTEM_PROMPT}\n",
"\n",
"Analyze this security threat:\n",
"{json.dumps(threat_data, indent=2)}\n",
"\n",
"Provide your analysis as JSON.\"\"\"\n",
" \n",
" try:\n",
" response = self.client.models.generate_content(\n",
" model=self.model_name,\n",
" contents=prompt,\n",
" config={\n",
" \"temperature\": 0.3,\n",
" \"max_output_tokens\": 1024,\n",
" }\n",
" )\n",
" \n",
" # Parse response\n",
" text = response.text\n",
" # Extract JSON from response\n",
" if '```json' in text:\n",
" text = text.split('```json')[1].split('```')[0]\n",
" elif '```' in text:\n",
" text = text.split('```')[1].split('```')[0]\n",
" \n",
" result = json.loads(text)\n",
" \n",
" return AgentDecision(\n",
" action=result.get('action', 'monitor'),\n",
" confidence=float(result.get('confidence', 0.5)),\n",
" reasoning=result.get('reasoning', 'Analysis pending'),\n",
" evidence=result.get('evidence', []),\n",
" risk_level=result.get('risk_level', 'medium'),\n",
" recommended_follow_up=result.get('recommended_follow_up', [])\n",
" )\n",
" except Exception as e:\n",
" print(f\" ⚠ Gemini error: {e}\")\n",
" return self._fallback_analyze(threat_data)\n",
" \n",
" def _fallback_analyze(self, threat_data: Dict) -> AgentDecision:\n",
" \"\"\"Rule-based fallback when Gemini unavailable\"\"\"\n",
" risk_score = threat_data.get('risk_score', 0.5)\n",
" indicators = threat_data.get('indicators', [])\n",
" \n",
" # Determine action based on risk score\n",
" if risk_score >= 0.8:\n",
" action = 'block'\n",
" risk_level = 'critical'\n",
" elif risk_score >= 0.6:\n",
" action = 'alert'\n",
" risk_level = 'high'\n",
" elif risk_score >= 0.4:\n",
" action = 'monitor'\n",
" risk_level = 'medium'\n",
" else:\n",
" action = 'allow'\n",
" risk_level = 'low'\n",
" \n",
" return AgentDecision(\n",
" action=action,\n",
" confidence=risk_score,\n",
" reasoning=f\"Risk score: {risk_score:.2f}. Indicators found: {len(indicators)}\",\n",
" evidence=[str(i) for i in indicators[:3]],\n",
" risk_level=risk_level,\n",
" recommended_follow_up=['Continue monitoring', 'Review threat logs']\n",
" )\n",
"\n",
"reasoning_engine = GeminiReasoningEngine()\n",
"print(\"✓ Gemini Reasoning Engine initialized\")"
]
},
{
"cell_type": "markdown",
"id": "a5f5a0a7",
"metadata": {},
"source": [
"## 4. Task Queue Manager"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "1f9ba4fe",
"metadata": {},
"outputs": [],
"source": [
"import uuid\n",
"from collections import deque\n",
"\n",
"class AgentTaskQueue:\n",
" \"\"\"\n",
" Manages agent task queue for autonomous operation.\n",
" Supports priority-based execution and task lifecycle.\n",
" \"\"\"\n",
" \n",
" def __init__(self, max_concurrent: int = 5):\n",
" self.pending = deque()\n",
" self.active = []\n",
" self.completed = []\n",
" self.failed = []\n",
" self.max_concurrent = max_concurrent\n",
" \n",
" def add_task(self, task_type: str, target: str, context: Dict,\n",
" priority: TaskPriority = TaskPriority.MEDIUM) -> AgentTask:\n",
" \"\"\"Add a new task to the queue\"\"\"\n",
" task = AgentTask(\n",
" task_id=str(uuid.uuid4())[:8],\n",
" task_type=task_type,\n",
" priority=priority.value,\n",
" target=target,\n",
" context=context,\n",
" created_at=time.time()\n",
" )\n",
" \n",
" self.pending.append(task)\n",
" self._reorder_queue()\n",
" \n",
" return task\n",
" \n",
" def _reorder_queue(self):\n",
" \"\"\"Reorder queue by priority\"\"\"\n",
" self.pending = deque(sorted(self.pending, \n",
" key=lambda t: (t.priority, -t.created_at),\n",
" reverse=True))\n",
" \n",
" def get_next_task(self) -> Optional[AgentTask]:\n",
" \"\"\"Get next task to execute\"\"\"\n",
" if len(self.active) >= self.max_concurrent:\n",
" return None\n",
" \n",
" if self.pending:\n",
" task = self.pending.popleft()\n",
" task.status = 'active'\n",
" self.active.append(task)\n",
" return task\n",
" \n",
" return None\n",
" \n",
" def complete_task(self, task_id: str, result: Dict, success: bool = True):\n",
" \"\"\"Mark task as completed\"\"\"\n",
" for i, task in enumerate(self.active):\n",
" if task.task_id == task_id:\n",
" task.status = 'completed' if success else 'failed'\n",
" task.result = result\n",
" \n",
" if success:\n",
" self.completed.append(task)\n",
" else:\n",
" self.failed.append(task)\n",
" \n",
" del self.active[i]\n",
" return\n",
" \n",
" def get_stats(self) -> Dict:\n",
" \"\"\"Get queue statistics\"\"\"\n",
" return {\n",
" 'pending': len(self.pending),\n",
" 'active': len(self.active),\n",
" 'completed': len(self.completed),\n",
" 'failed': len(self.failed),\n",
" 'total': len(self.pending) + len(self.active) + len(self.completed) + len(self.failed)\n",
" }\n",
"\n",
"task_queue = AgentTaskQueue()\n",
"print(\"✓ Task Queue Manager initialized\")"
]
},
{
"cell_type": "markdown",
"id": "1a56cbf3",
"metadata": {},
"source": [
"## 5. Autonomous Agent Orchestrator"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "3a79fff8",
"metadata": {},
"outputs": [],
"source": [
"class CyberForgeAgent:\n",
" \"\"\"\n",
" Main autonomous agent for CyberForge.\n",
" Orchestrates threat detection, analysis, and response.\n",
" \"\"\"\n",
" \n",
" def __init__(self):\n",
" self.scoring_engine = DecisionScoringEngine()\n",
" self.reasoning_engine = GeminiReasoningEngine()\n",
" self.task_queue = AgentTaskQueue()\n",
" self.action_history = []\n",
" \n",
" def analyze_website(self, url: str, scraped_data: Dict) -> AgentDecision:\n",
" \"\"\"Analyze a website for threats\"\"\"\n",
" # Extract indicators\n",
" indicators = self._extract_indicators(scraped_data)\n",
" \n",
" # Calculate threat score\n",
" score, risk_level = self.scoring_engine.calculate_threat_score(indicators)\n",
" \n",
" # Get AI reasoning\n",
" threat_data = {\n",
" 'url': url,\n",
" 'risk_score': score,\n",
" 'indicators': indicators,\n",
" 'security_report': scraped_data.get('security_report', {})\n",
" }\n",
" \n",
" decision = self.reasoning_engine.analyze_threat(threat_data)\n",
" \n",
" # Log decision\n",
" self._log_action(url, decision)\n",
" \n",
" return decision\n",
" \n",
" def _extract_indicators(self, data: Dict) -> List[Dict]:\n",
" \"\"\"Extract threat indicators from scraped data\"\"\"\n",
" indicators = []\n",
" \n",
" # Check security report\n",
" security = data.get('security_report', {})\n",
" if not security.get('is_https', True):\n",
" indicators.append({\n",
" 'type': 'insecure_protocol',\n",
" 'severity': 'medium',\n",
" 'confidence': 0.9,\n",
" 'evidence_type': 'signature_match'\n",
" })\n",
" \n",
" if security.get('mixed_content', False):\n",
" indicators.append({\n",
" 'type': 'mixed_content',\n",
" 'severity': 'medium',\n",
" 'confidence': 0.85,\n",
" 'evidence_type': 'signature_match'\n",
" })\n",
" \n",
" # Check console errors\n",
" console_logs = data.get('console_logs', [])\n",
" errors = [log for log in console_logs if log.get('level') == 'error']\n",
" if len(errors) > 5:\n",
" indicators.append({\n",
" 'type': 'excessive_errors',\n",
" 'severity': 'low',\n",
" 'confidence': 0.6,\n",
" 'evidence_type': 'behavioral_pattern'\n",
" })\n",
" \n",
" # Check for suspicious network requests\n",
" requests = data.get('network_requests', [])\n",
" suspicious_domains = ['malware', 'phishing', 'hack', 'tracker']\n",
" for req in requests:\n",
" url = req.get('url', '').lower()\n",
" if any(s in url for s in suspicious_domains):\n",
" indicators.append({\n",
" 'type': 'suspicious_request',\n",
" 'severity': 'high',\n",
" 'confidence': 0.75,\n",
" 'evidence_type': 'signature_match',\n",
" 'details': url[:100]\n",
" })\n",
" \n",
" return indicators\n",
" \n",
" def _log_action(self, target: str, decision: AgentDecision):\n",
" \"\"\"Log agent action for audit trail\"\"\"\n",
" self.action_history.append({\n",
" 'timestamp': time.time(),\n",
" 'target': target,\n",
" 'action': decision.action,\n",
" 'confidence': decision.confidence,\n",
" 'risk_level': decision.risk_level\n",
" })\n",
" \n",
" def get_action_summary(self) -> Dict:\n",
" \"\"\"Get summary of agent actions\"\"\"\n",
" if not self.action_history:\n",
" return {'total_actions': 0}\n",
" \n",
" actions = [a['action'] for a in self.action_history]\n",
" return {\n",
" 'total_actions': len(self.action_history),\n",
" 'action_counts': {a: actions.count(a) for a in set(actions)},\n",
" 'avg_confidence': np.mean([a['confidence'] for a in self.action_history])\n",
" }\n",
"\n",
"agent = CyberForgeAgent()\n",
"print(\"✓ CyberForge Agent initialized\")"
]
},
{
"cell_type": "markdown",
"id": "7b7ad27d",
"metadata": {},
"source": [
"## 6. Test Agent Decision Making"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a377b332",
"metadata": {},
"outputs": [],
"source": [
"# Test with sample threat data\n",
"test_data = {\n",
" 'url': 'https://suspicious-login.example.com/verify',\n",
" 'security_report': {\n",
" 'is_https': True,\n",
" 'mixed_content': True,\n",
" 'insecure_cookies': True\n",
" },\n",
" 'console_logs': [\n",
" {'level': 'error', 'message': 'CORS policy violation'},\n",
" {'level': 'error', 'message': 'Failed to load resource'},\n",
" ],\n",
" 'network_requests': [\n",
" {'url': 'https://tracker.malicious.com/collect', 'type': 'xhr'},\n",
" {'url': 'https://cdn.example.com/app.js', 'type': 'script'}\n",
" ]\n",
"}\n",
"\n",
"print(\"Testing agent analysis...\\n\")\n",
"decision = agent.analyze_website(test_data['url'], test_data)\n",
"\n",
"print(f\"Decision: {decision.action}\")\n",
"print(f\"Confidence: {decision.confidence:.2%}\")\n",
"print(f\"Risk Level: {decision.risk_level}\")\n",
"print(f\"\\nReasoning: {decision.reasoning}\")\n",
"print(f\"\\nEvidence:\")\n",
"for e in decision.evidence[:3]:\n",
" print(f\" - {e}\")\n",
"print(f\"\\nRecommended Follow-up:\")\n",
"for r in decision.recommended_follow_up[:3]:\n",
" print(f\" - {r}\")"
]
},
{
"cell_type": "markdown",
"id": "e68ef74b",
"metadata": {},
"source": [
"## 7. Save Agent Configuration"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "cde8619f",
"metadata": {},
"outputs": [],
"source": [
"# Save agent configuration and modules\n",
"agent_config = {\n",
" 'version': '1.0.0',\n",
" 'confidence_threshold': 0.7,\n",
" 'max_concurrent_tasks': 5,\n",
" 'severity_weights': DecisionScoringEngine.SEVERITY_WEIGHTS,\n",
" 'evidence_weights': DecisionScoringEngine.EVIDENCE_WEIGHTS,\n",
" 'task_priorities': {p.name: p.value for p in TaskPriority},\n",
" 'gemini_model': CONFIG.get('gemini_model', 'gemini-2.5-flash')\n",
"}\n",
"\n",
"config_path = AGENT_DIR / \"agent_config.json\"\n",
"with open(config_path, 'w') as f:\n",
" json.dump(agent_config, f, indent=2)\n",
"\n",
"print(f\"✓ Agent config saved to: {config_path}\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "43b960cd",
"metadata": {},
"outputs": [],
"source": [
"# Save agent module for backend import\n",
"agent_module = '''\n",
"\"\"\"CyberForge Agent Intelligence Module\"\"\"\n",
"\n",
"import json\n",
"import time\n",
"import numpy as np\n",
"from pathlib import Path\n",
"from dataclasses import dataclass, asdict\n",
"from typing import Dict, List, Any, Optional\n",
"\n",
"@dataclass\n",
"class AgentDecision:\n",
" action: str\n",
" confidence: float\n",
" reasoning: str\n",
" evidence: List[str]\n",
" risk_level: str\n",
" recommended_follow_up: List[str]\n",
" \n",
" def to_dict(self):\n",
" return asdict(self)\n",
"\n",
"class DecisionEngine:\n",
" SEVERITY_WEIGHTS = {\"critical\": 1.0, \"high\": 0.8, \"medium\": 0.5, \"low\": 0.3, \"info\": 0.1}\n",
" \n",
" def calculate_threat_score(self, indicators: List[Dict]) -> tuple:\n",
" if not indicators:\n",
" return 0.0, \"low\"\n",
" scores = [i.get(\"confidence\", 0.5) * self.SEVERITY_WEIGHTS.get(i.get(\"severity\", \"low\"), 0.3) \n",
" for i in indicators]\n",
" score = sum(scores) / len(scores) if scores else 0\n",
" risk = \"critical\" if score >= 0.8 else \"high\" if score >= 0.6 else \"medium\" if score >= 0.4 else \"low\"\n",
" return score, risk\n",
"\n",
"class CyberForgeAgent:\n",
" def __init__(self):\n",
" self.engine = DecisionEngine()\n",
" \n",
" def analyze(self, url: str, data: Dict) -> Dict:\n",
" indicators = self._extract_indicators(data)\n",
" score, risk = self.engine.calculate_threat_score(indicators)\n",
" action = \"block\" if score >= 0.8 else \"alert\" if score >= 0.6 else \"monitor\" if score >= 0.4 else \"allow\"\n",
" \n",
" return AgentDecision(\n",
" action=action,\n",
" confidence=score,\n",
" reasoning=f\"Threat score: {score:.2f}. {len(indicators)} indicators found.\",\n",
" evidence=[str(i) for i in indicators[:3]],\n",
" risk_level=risk,\n",
" recommended_follow_up=[\"Continue monitoring\"]\n",
" ).to_dict()\n",
" \n",
" def _extract_indicators(self, data: Dict) -> List[Dict]:\n",
" indicators = []\n",
" sec = data.get(\"security_report\", {})\n",
" if not sec.get(\"is_https\", True):\n",
" indicators.append({\"type\": \"insecure\", \"severity\": \"medium\", \"confidence\": 0.9})\n",
" if sec.get(\"mixed_content\"):\n",
" indicators.append({\"type\": \"mixed_content\", \"severity\": \"medium\", \"confidence\": 0.85})\n",
" return indicators\n",
"'''\n",
"\n",
"module_path = AGENT_DIR / \"cyberforge_agent.py\"\n",
"with open(module_path, 'w') as f:\n",
" f.write(agent_module)\n",
"\n",
"print(f\"✓ Agent module saved to: {module_path}\")"
]
},
{
"cell_type": "markdown",
"id": "31978536",
"metadata": {},
"source": [
"## 8. Summary"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "d6e9505c",
"metadata": {},
"outputs": [],
"source": [
"print(\"\\n\" + \"=\" * 60)\n",
"print(\"AGENT INTELLIGENCE COMPLETE\")\n",
"print(\"=\" * 60)\n",
"\n",
"print(f\"\"\"\n",
"🤖 Agent Capabilities:\n",
" - Decision Scoring: Weighted threat assessment\n",
" - Gemini Integration: AI-powered reasoning\n",
" - Task Queue: Priority-based execution\n",
" - Action History: Full audit trail\n",
"\n",
"📊 Test Results:\n",
" - Action: {decision.action}\n",
" - Confidence: {decision.confidence:.2%}\n",
" - Risk Level: {decision.risk_level}\n",
"\n",
"📁 Output Files:\n",
" - Config: {AGENT_DIR}/agent_config.json\n",
" - Module: {AGENT_DIR}/cyberforge_agent.py\n",
"\n",
"Next step:\n",
" → 05_model_validation.ipynb\n",
"\"\"\")\n",
"print(\"=\" * 60)"
]
}
],
"metadata": {
"language_info": {
"name": "python"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
|