Mitsu
Meet Mitsu, your chibi red-team operator. Connect your AI provider, set the scope, and she can take an authorized pentest from recon and attack-surface mapping
What is Mitsu?
Meet Mitsu, your chibi red-team operator. Connect your AI provider, set the scope, and she can take an authorized pentest from recon and attack-surface mapping through vulnerability discovery, > exploitation, and post-exploitation, armored with 165+ MCP-connected tools and live session intel. Guide every step or let her pwn autonomously. ( ◠‿◠)
Overview of Mitsu
What it is
Mitsu is an agentic framework client designed for authorized penetration-testing workflows. It connects to a chosen AI model and, given an in-scope target and objective, can carry out reconnaissance, enumerate hosts and services, research vulnerabilities and exploits, validate findings, and continue into post-exploitation. Its MCP server connects the selected model to over 165 Red Team Armory tools, external MCP servers, and custom command wrappers. Each session maintains its own Intel, including targets, IPs, domains, services, vulnerabilities, findings, accounts, evidence, and observations. The session Intel can be turned into an interactive attack-surface map or an assessment report. Mitsu is open source under the MIT license and requires Python 3.10 or newer.
Key features
- Connects to model providers including OpenAI, Anthropic, OpenRouter, Ollama, DeepSeek, Moonshot (Kimi), NVIDIA NIM, and Cloudflare Models, plus custom OpenAI- or Anthropic-compatible endpoints.
- Provides an MCP server that exposes a catalog of 165+ Red Team Armory tools across 17 categories, plus custom CLI wrappers and external MCP servers.
- Maintains persistent engagement Intel for targets, findings, services, vulnerabilities, domains, IPs, accounts, people, and working observations.
- Offers three operating modes: Recon (passive, operator-led), Guided (operator steers with action approval), and AutoPwn (autonomous evidence-led next steps within authorized scope).
- Supports engagement-scoped sessions that keep conversation, Intel, artifacts, and working directory together; sessions can be resumed, searched, rewound, or exported.
- Includes vulnerability research that combines session Intel with configured search engines and reference databases.
Use cases
Mitsu is intended for authorized penetration-testing engagements. Use cases include performing reconnaissance on a target, enumerating hosts and services, researching and validating vulnerabilities, conducting exploitation and post-exploitation activities, and generating assessment reports or attack-surface maps from session Intel. It can be used in a guided manner where the operator directs each step, or in an autonomous mode where Mitsu chooses evidence-led next steps within the defined scope.
Getting started
Install Mitsu with Python 3.10 or newer. Using pip: sh pip install mitsu mitsu Install from GitHub: sh git clone https://github.com/ARCANGEL0/Mitsu.git cd Mitsu ./scripts/install.sh Install from a local checkout: sh uv tool install . mitsu On first launch, the setup walks through configuring a model provider, choosing a model, configuring web-search providers, selecting Red Team Armory tools, and choosing Mitsu’s companion.
License and pricing
Mitsu is licensed under the MIT license. It is open source and free to use. The license is stated in the repository as MIT. Pricing type is open source. No other pricing information is provided.
Based on the project's public README and repository metadata.