LLMtary is an autonomous penetration testing application that leverages large language models (LLMs) to automate the entire security assessment lifecycle. Designed for Windows, macOS, and Linux, it streamlines processes from reconnaissance and vulnerability analysis to active exploit validation and professional report generation. Unlike traditional tools that merely suggest potential vulnerabilities, LLMtary executes real commands to confirm or refute each finding, ensuring accurate and actionable results.
Key Features and Functionality:
- Autonomous Reconnaissance: Utilizes LLM-guided engines to perform discovery tasks such as port scans, service banner identification, DNS enumeration, and WAF detection, compiling enriched target data for analysis.
- Two-Phase Analysis Pipeline: Phase 1 focuses on CVE matching, DNS/OSINT, and network service analysis. Phase 2 conducts targeted assessments on web applications, Active Directory, and specific technologies, building upon Phase 1 insights.
- Active Exploit Testing Loop: Employs an iterative process where the LLM plans, executes, evaluates, and adapts real shell commands against targets to confirm or dismiss vulnerabilities.
- Attack Chain Reasoning: When multiple vulnerabilities are confirmed, the tool identifies and maps out multi-step attack paths, enhancing understanding of potential security breaches.
- Credential Management: Collects discovered credentials session-wide, automatically reusing them for subsequent targets and triggering authenticated re-analysis for deeper insights.
- Safety Controls: Incorporates hard blocklists to prevent execution of dangerous commands and offers a command approval mode for user review before execution, ensuring operational safety.
- Post-Exploitation Enumeration: Upon confirming remote code execution or significant access, it automatically enumerates users, credentials, network interfaces, running services, and privilege escalation paths.
- Professional Reporting: Generates comprehensive reports in HTML, Markdown, or CSV formats, featuring AI-assisted executive summaries, CVSS metadata, discovered credentials, and detailed attack chain narratives.
Primary Value and User Solutions:
LLMtary addresses the need for efficient, accurate, and comprehensive penetration testing by automating complex tasks traditionally performed manually. By integrating LLMs, it reduces the time and expertise required to conduct thorough security assessments, making it accessible to organizations of varying sizes. Its ability to confirm vulnerabilities through real command execution minimizes false positives, providing security teams with reliable data to prioritize and remediate threats effectively. The tool's structured approach mirrors real-world engagement workflows, ensuring that each phase enriches the next, leading to a holistic understanding of an organization's security posture.
