Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/SanMuzZzZz/LuaN1aoAgent/llms.txt

Use this file to discover all available pages before exploring further.

Quick Start

Get the agent running against your first target in minutes.

Core Concepts

Understand the P-E-R architecture and dual-graph reasoning engine.

Configuration

Configure LLM providers, models, and execution parameters.

Tool Reference

Explore every MCP tool available to the agent during execution.

​
What is LuaN1aoAgent?

LuaN1aoAgent (鸾鸟) is a next-generation autonomous penetration testing agent. Unlike traditional scanners that rely on predefined rules, LuaN1ao simulates the cognitive process of a human security expert — planning attack paths, building evidence-driven hypotheses, executing targeted actions, and learning from failures in a continuous loop. The agent is built around three core innovations:

P-E-R Architecture

Planner, Executor, and Reflector agents collaborate via an event bus, each focusing on their core role to eliminate the “split personality” problem of single-agent systems.

Causal Graph Reasoning

Every hypothesis requires explicit evidence. The agent builds rigorous Evidence → Hypothesis → Vulnerability → Exploit chains with confidence scores to prevent hallucinated attacks.

Plan-on-Graph (PoG)

Tasks are modeled as dynamically evolving Directed Acyclic Graphs (DAGs), enabling parallel execution, real-time adaptation, and topological dependency management.

​
Key metrics

MetricValue
XBOW Benchmark success rate90.4%
Median exploit cost$0.09
LLM compatibilityOpenAI, DeepSeek, Claude, any OpenAI-compatible API
Python requirement3.10+

​
How it works

User Goal → Planner (DAG planning) → Executor (tool invocation) → Reflector (audit + causal graph)
                    ↑                                                          |
                    └──────────── Intelligence feedback loop ──────────────────┘
The agent runs as two separate processes:
  • Web Server (python -m web.server) — persistent dashboard at localhost:8088
  • Agent (python agent.py --goal "...") — executes the task, writes to SQLite, exits when done
LuaN1aoAgent contains high-privilege tools including shell_exec and python_exec. Always run in a Docker container or virtual machine to protect your host system. Never test against systems you do not own or have explicit written authorization to test.

​
Get started

1

Install dependencies

git clone https://github.com/SanMuzZzZz/LuaN1aoAgent.git
cd LuaN1aoAgent
python3 -m venv venv && source venv/bin/activate
pip install -r requirements.txt
2

Configure your LLM

cp .env.example .env
# Edit .env with your LLM_API_KEY and LLM_API_BASE_URL
3

Build the knowledge base

git clone https://github.com/swisskyrepo/PayloadsAllTheThings knowledge_base/PayloadsAllTheThings
cd rag && python -m rag_kdprepare
4

Start the web server and run a task

# Terminal 1
python -m web.server

# Terminal 2
python agent.py --goal "Perform web security testing on http://testphp.vulnweb.com" --task-name "demo"
View the full Quick Start guide for more detail, or explore Core Concepts to understand the agent’s reasoning engine.

Build docs developers (and LLMs) love

Get started for freeTalk to us