RedAmon is an open-source, self-hosted AI penetration testing framework that maps your attack surface into a knowledge graph, then autonomously probes and exploits it from a Kali Linux sandbox, with human approval gates before critical actions. When it finds a weakness, it goes further still: it triages the finding, writes a code fix, and opens a pull request on your repository.
RedAmon is developed in the open on GitHub, where it has earned more than 3,100 stars under the MIT license, and it remains actively maintained, with the latest updates in October 2026. The framework chains reconnaissance, exploitation, and post-exploitation into a single pipeline, storing the target’s infrastructure as a graph (backed by Neo4j) so the agent can reason about relationships between hosts, services, and vulnerabilities instead of scanning blindly. On the XBOW web-security benchmark it reports solving 101 of 104 challenges fully black-box, with complete raw agent sessions published for audit. RedAmon speaks the Model Context Protocol in both directions: you can plug any MCP server into it as a tool, or drive RedAmon itself from Claude Code or your own agent. It is intended strictly for authorized security testing on systems you own or have written permission to test.
RedAmon is completely free and open source under the MIT license. You self-host it and supply your own LLM access and testing infrastructure, so there are no subscription fees or per-scan charges.
RedAmon suits penetration testers, red teams, bug bounty hunters, and application security engineers who want an autonomous agent to handle repetitive recon and exploitation work while keeping a human in control of consequential steps. It is a strong fit for teams that already live in Kali Linux and want findings delivered as fixes, not just reports.