PentestAgent is an AI agent framework for black-box security testing, enabling three modes of operation and MCP integration. It provides built-in tools, playbooks, and Docker options for running pentesting tasks.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
PentestAgent is an AI Penetration Testing framework that supports black-box security testing workflows, including bug bounty, red-team, and penetration testing. It includes playbooks for attack workflows and a set of built-in tools and MCP (Model Context Protocol) integration for extensibility.
How it works
The project exposes a TUI with modes: Assist, Agent, Crew, and Interact for task execution and orchestration. It includes an MCP capability to consume external MCP servers or to expose PentestAgent as an MCP server via STDIO or SSE transports. It provides an internal toolset (e.g., terminal, browser, notes, web_search) and supports an agent self-spawning feature for hierarchical workflows. It also offers an MCP Server Tools Reference for task execution, inspection, and memory/observability tooling. Playbooks define structured approaches for security assessments and can be run with a command that specifies a target and playbook.
Getting started
Getting started steps are provided under Install and Run sections: clone the repository, set up a Python virtual environment, install dependencies, and configure environment variables for provider keys. The README includes commands like:
# Clone
git clone https://github.com/GH05TCREW/pentestagent.git
cd pentestagent
# Setup (creates venv, installs deps)
.\scripts\setup.ps1 # Windows
./scripts/setup.sh # Linux/macOS
# Or manual
python -m venv venv
.\venv\Scripts\Activate.ps1 # Windows
source venv/bin/activate # Linux/macOS
pip install -e ".[all]"
playwright install chromium # Required for browser tool
Configure environment variables in a .env file, including ANTHROPIC/OpenAI API keys and PENTESTAGENT_MODEL settings, as shown in the Configure section. Run the UI with:
pentestagent # Launch TUI
pentestagent -t 192.168.1.1 # Launch with target
pentestagent tui --docker # Run tools in Docker container
Recent releases
The latest release entry states 0.2.0 in the badges, but the RELEASES section lists none as latest. The README presents a Version badge showing 0.2.0.
Traction
GitHub statistics present: 2867 stars and 575 forks. Open issues: 23. Language: Python. License: MIT. Created: 2025-05-15. Last push: 2026-08-04.
Behind the repo
The repository name indicates a single project under GH05TCREW; no external startup/company linkage is provided beyond the project metadata.
Caveats
License: MIT. Requirements include Python 3.10+ and an API key for LiteLLM-supported providers. Docker runs require Docker installation. No explicit age beyond created date; latest push is 2026-08-04.