PentestGPT
β Text source-checked Nov 9, 2026Automated Penetration Testing Agentic Framework Powered by Large Language Models
Source-checked against the vendor's site Β· rebuilt from the vendor's own pagesProfile updated Link checked How we rate
In short
PentestGPT is a free, MIT-licensed AI agent for automated penetration testing. It uses LLMs to handle recon, exploitation, and reporting for CTFs and security audits.
PentestGPT is an AI-powered autonomous penetration testing agent. The agent operates through a multi-stage pipeline that feeds findings from each phase into the next, covering recon, exploit, and walkthrough for CTFs, or asset discovery, vulnerability identification, and report generation for pentests.
In autonomous mode, PentestGPT drives Claude Code or Codex to execute tools and reason without human intervention. An interactive legacy mode supports a wider set of LLM providers including OpenAI, Anthropic, Google Gemini, DeepSeek, xAI, Qwen, Moonshot, and local Ollama.
The project is distributed under the MIT License. The tool was published at USENIX Security 2024.
According to PentestGPT: PentestGPT is an AI-powered autonomous penetration testing agent. Its pages list this strength: Autonomous agent drives Claude Code or Codex without human intervention.
β The AI Tool Bible editorial team
Pros
- β Autonomous agent drives Claude Code or Codex without human intervention
- β Multi-stage pipeline feeds each stage's findings into the next
- β Session persistence allows saving and resuming penetration testing sessions
- β Extensible architecture with a clean, modular design
- β Supports multiple LLM providers including OpenAI, Anthropic, and local Ollama
Facts
- Platforms
- cli (not source-checked)
- Pricing
- free - MIT license (not source-checked)
- Open source
- Yes (license: MIT)
- GitHub stars
- 16,355 (checked Nov 10, 2026) (not source-checked)
- Last GitHub push
- Jul 14, 2026 (not source-checked)
- First commit
- Feb 2023 (not source-checked)
Where this tool sits
Use cases
Frequently asked
- How much does PentestGPT cost?
- PentestGPT is free to use. The project is distributed under the MIT License, allowing for open-source access and distribution without licensing fees.
- Which AI models does PentestGPT support?
- In autonomous mode, it drives Claude Code or Codex. Interactive legacy mode supports OpenAI, Anthropic, Google Gemini, DeepSeek, xAI, Qwen, Moonshot, and local Ollama models.
- What specific tasks can PentestGPT automate?
- It automates penetration testing, asset discovery, vulnerability identification, and security report generation. It also supports live walkthroughs of AI reasoning for CTFs and pentests.
- Is PentestGPT suitable for beginners?
- The documentation does not specify a learning curve or beginner suitability. It is described as an autonomous agent that operates through a multi-stage pipeline for complex security tasks.
- Where was PentestGPT published?
- The tool was published at USENIX Security 2024. It is an AI-powered autonomous penetration testing agent designed to feed findings from each phase into the next.