samugit83/redamon

What it solves

RedAmon is an autonomous AI framework designed to automate the entire security testing lifecycle. It bridges the gap between offensive security (finding vulnerabilities) and defensive remediation (fixing them) by chaining reconnaissance, exploitation, and automated code fixes into a single pipeline.

How it works

The system operates as an end-to-end pipeline consisting of several key phases:

1. Reconnaissance: A 6-phase engine launches multiple security tools (like Nmap, Nuclei, and OpenVAS) in parallel to map the attack surface. Results are merged into a Neo4j knowledge graph for structured, natural-language querying.
2. Exploitation: An autonomous AI agent uses the knowledge graph to validate CVE exploitability, test credential policies (via Hydra), and map lateral movement paths.
3. Triage: The CypherFix agent correlates and deduplicates findings, ranking them by exploitability.
4. Remediation: A CodeFix agent clones the target repository, uses code-aware tools to implement fixes, and opens a GitHub pull request for human review.

Who it’s for

It is intended for security researchers, penetration testers, and developers who need to automate vulnerability discovery and the subsequent patching process in authorized environments.

Highlights

• Full Lifecycle Automation: Moves from initial packet capture to merged pull requests.
• Knowledge Graph Integration: Uses Neo4j to maintain a living map of the attack surface.
• Multi-Agent Parallelism: Employs "Fireteam" parallel specialists to test different vulnerability paths simultaneously.
• Extensive Tooling: Integrates 100+ security tools and supports 400+ AI models via providers like OpenAI, Anthropic, and local options via Ollama or vLLM.
• Auto-Remediation: Automatically implements code fixes and submits them via GitHub PRs.