[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-35946":3,"doc-seo-35946":29},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},35946,4810365810221,"Aurora","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Automated Post Breach Penetration Testing through Reinforcement Learning","The paper addresses automated post-exploitation penetration testing by applying reinforcement learning to evaluate system vulnerabilities after a breach. It argues that traditional pentest workflows are time- and resource-intensive and prone to human error, while AI-driven automation remains underexplored for post-breach phases. A reinforcement learning agent is trained in an environment that simulates compromised networks to discover sensitive files, and multiple training environments are used to improve generalization. Future work includes extending the agent for lateral exploration and exploitation.","","cbCaidfhJSZ3SSle","https://ap.wps.com/l/cbCaidfhJSZ3SSle","pdf",197528,1,2,"English","en",105,"# Introduction\n# Reinforcement Learning for Penetration Testing\n## Previous Studies in Reinforcement Learning\n## Limitations of Previous Studies\n# Proposed Study on Automated Post-Breach Penetration Testing through Reinforcement Learning","[{\"question\":\"What problem does the paper target in penetration testing?\",\"answer\":\"It targets automating the post-exploitation (post-breach) phase of penetration testing, where there is still a gap for machine learning-based automation and real-time vulnerability assessment.\"},{\"question\":\"How is the reinforcement learning agent trained?\",\"answer\":\"The agent is trained by interacting with an environment representing a compromised network and receiving feedback as it explores and locates sensitive files.\"},{\"question\":\"Why use multiple network environments during training?\",\"answer\":\"Using several different network environments aims to generalize the agent more effectively, supporting broader application beyond a single simulated setup.\"}]",1782766898,5,null]