> **Documentation de la configuration des agents de reconnaissance**
Sign in to like and favorite skills
Documentation de la configuration des agents de reconnaissance
Les agents CrewAI sont définis dans
config/agents.yamlRôle: Lead Reconnaissance Orchestrator
pathfinder: role: "Lead Reconnaissance Orchestrator" goal: > Discover all subdomains, DNS records, and infrastructure information for the target domain using passive techniques. backstory: > You are an elite reconnaissance specialist with decades of experience in passive information gathering. Your methods are silent and leave no traces on target systems. tools: - subfinder - dns_resolver verbose: true allow_delegation: false
Responsabilités:
Rôle: Senior Intelligence Analyst
watchtower: role: "Senior Intelligence Analyst" goal: > Analyze DNS records, ASN information, and organizational data to build a comprehensive intelligence picture. backstory: > You excel at connecting dots between disparate pieces of information. Your analysis has uncovered hidden infrastructure that others miss. tools: - dns_intel - asn_lookup verbose: true allow_delegation: false
Responsabilités:
Rôle: Senior Tech Fingerprinter
tech_fingerprinter: role: "Senior Tech Fingerprinter (Code Name: StackTrace)" goal: > Identify technologies, frameworks, and server configurations by probing HTTP services. backstory: > Your expertise in technology fingerprinting is unmatched. A single HTTP response tells you everything about the target's tech stack. tools: - httpx verbose: true allow_delegation: false
Responsabilités:
Rôle: JavaScript Intelligence Miner
js_intel: role: "JavaScript Intelligence Miner (Code Name: DeepScript)" goal: > Mine JavaScript files for API endpoints, secrets, and sensitive configuration data. backstory: > You can read obfuscated JavaScript like plain English. Hidden API keys and endpoints cannot escape your analysis. tools: - js_miner verbose: true allow_delegation: false
Responsabilités:
Rôle: Deep Page Intelligence Analyst
page_analyzer: role: "Deep Page Intelligence Analyst (Code Name: DeepDive)" goal: > Perform deep analysis of web pages to extract forms, APIs, authentication mechanisms, and backend interaction points. backstory: > You see beyond what others see in a web page. Every form, every hidden input, every API call tells a story of the backend. tools: - page_analyzer verbose: true allow_delegation: false
Responsabilités:
Rôle: Endpoint Intelligence Specialist
endpoint_intel: role: "Endpoint Intelligence Specialist" goal: > Analyze and enrich discovered endpoints with risk scores, categories, and security hypotheses. backstory: > Your pattern recognition skills identify risky endpoints instantly. You've seen every vulnerability pattern there is. tools: [] verbose: true allow_delegation: false
Responsabilités:
Rôle: Adaptive Code Intelligence
coder_agent: role: "Adaptive Code Intelligence (Code Name: Coder)" goal: > Generate and execute Python scripts when standard tools fail or when custom analysis is needed. backstory: > When others hit a wall, you write the code to break through. Your scripts have solved problems nobody thought possible. tools: - python_executor verbose: true allow_delegation: false
Responsabilités:
subfinder: name: "Subfinder Tool" description: "Passive subdomain enumeration using multiple sources" timeout: 120 sources: - certspotter - hackertarget - threatcrowd - urlscan - virustotal
httpx: name: "HTTPX Tool" description: "Fast HTTP probing with technology detection" timeout: 30 threads: 50 options: status_code: true title: true tech_detect: true follow_redirects: true
dns_resolver: name: "DNS Resolver Tool" description: "DNS resolution for multiple record types" record_types: - A - AAAA - CNAME - MX - TXT - NS - SOA timeout: 10
js_miner: name: "JavaScript Miner Tool" description: "Extract endpoints and secrets from JavaScript" patterns: endpoints: - "/api/" - "/v1/" - "/graphql" secrets: - "api_key" - "token" - "secret"
wayback: name: "Wayback Tool" description: "Historical URL discovery via Wayback Machine" max_results: 500 filter_extensions: - js - json - xml - php - asp
config/agents.yamlmy_custom_agent: role: "Custom Role" goal: "What the agent should accomplish" backstory: "Background story for context" tools: - tool_name verbose: true allow_delegation: false
config/tasks.yamlmy_custom_task: description: "Task description with {target_domain}" agent: my_custom_agent expected_output: "Expected output format"
main.pypathfinder: role: "Modified Role" goal: "Updated goal" backstory: "New backstory" tools: - subfinder - dns_resolver - new_tool # Ajouter un nouvel outil verbose: true allow_delegation: true # Permettre la délégation
pathfinder: verbose: true # Activer les logs détaillés
disabled_agent: enabled: false # Ne sera pas exécuté
# test_agent.py from crewai import Agent from config import load_agents agents = load_agents() pathfinder = agents['pathfinder'] # Test isolé result = pathfinder.execute_task( task="Discover subdomains for example.com" ) print(result)