PentestGPT

PentestGPT is an automated penetration testing agentic framework powered by large language models.

PentestGPT is an AI-powered autonomous penetration testing agent designed for security testing and capture-the-flag challenges. It operates through a multi-stage pipeline, moving from reconnaissance to exploitation and reporting. The tool supports an autonomous mode using Claude Code or Codex, as well as a modernized interactive legacy mode compatible with OpenAI, Anthropic, Google Gemini, DeepSeek, xAI, Qwen, Moonshot, and local Ollama models. Users can run the tool via CLI or in a self-contained Docker image with session persistence. It is built for security professionals and researchers conducting authorized testing.

PentestGPT at a glance

Best for: Software engineer
Runs on: CLI, self-hosted, open source
Pricing: Free
Released: February 2023
Category: AI coding tool
Website: github.com/GreyDGL/PentestGPT

Key features

Multi-stage pipeline for automated security testing and CTFs
Autonomous agent driving Claude Code and Codex CLIs
Session persistence to save and resume penetration testing sessions
Interactive multi-LLM mode maintaining a Pentesting Task Tree
Live walkthrough and real-time activity updates
Support for OpenAI, Anthropic, Google Gemini, DeepSeek, xAI, Qwen, Moonshot, and Ollama
Multi-category support including web, crypto, reversing, forensics, PWN, and privilege escalation

Use it for

Perform automated penetration testing and asset discovery
Solve capture-the-flag (CTF) security challenges
Execute multi-stage vulnerability identification and reporting
Run interactive security analysis with multiple cooperating LLM sessions

Alternatives to PentestGPT

Autonomous Pentest Agent — Dual-graph architecture for asset synchronization and multi-step attack planning.
PentAGI — It autonomously manages penetration testing workflows within a secure Docker sandbox.
Pentest AI — Eliminates false positives by reproducing every exploit three times.
PentestAgent — It provides autonomous multi-agent penetration testing and Docker tool isolation.
ARTEX — Runs autonomous multi-agent penetration testing with a dual-graph architecture.
VulnClaw — It uses LLM agents to autonomously execute end-to-end security testing loops.
pstack — Improves AI coding agent outcomes with rigorous workflows.
Artex Global — Autonomous multi-agent penetration testing with multilingual output support.
PentesterFlow — Pentest agent that runs on local models and writes findings
ARTEX — Manages autonomous penetration testing tasks and tracks vulnerability progress.
CyberStrike — Autonomous AI agents execute recon, web, network, and cloud penetration testing.
Oracle — Provides a single entry point for multiple pro AI agents and session management.

All PentestGPT alternatives