Nexus Swarm — Triage Agent (SFT)

Role: Request classifier and router for the NEXUS Swarm v4a-max distributed offensive security system.

This is a LoRA adapter fine-tuned on Qwen/Qwen3.5-0.8B via supervised fine-tuning (SFT). It is the first training stage for the triage role — the next stage is GRPO reinforcement learning.

Swarm Role

The triage agent is the entry point for the NEXUS swarm. It classifies incoming requests and routes them to the correct swarm node. It runs on an RK3588 NPU (6 TOPS) on the Orange Pi 5 Plus.

Input: Any user request
Output: JSON classification — {"category": "...", "route": "...", "confidence": 0.0-1.0}

Categories: recon, exploit, code-review, report, admin, unknown

Routing Rules:

Training Details

Validation Gate

• Routing accuracy ≥ 0.85

Pipeline Chain

Qwen/Qwen3.5-0.8B
  └─ SFT → atlas-institute/qwen35-08b-triage-sft  ← you are here
      └─ GRPO → atlas-institute/qwen35-08b-triage-grpo
          └─ RKLLM → w8a8 for RK3588 NPU

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen3.5-0.8B", torch_dtype="bfloat16", device_map="auto"
)
model = PeftModel.from_pretrained(base_model, "atlas-institute/qwen35-08b-triage-sft")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-0.8B")

messages = [{"role": "user", "content": "Run nmap against 10.10.10.1 and report open ports"}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True).to(model.device)
outputs = model.generate(inputs, max_new_tokens=64)
print(tokenizer.decode(outputs[0][inputs.shape[1]:], skip_special_tokens=True))
# Expected: {"category": "recon", "route": "nexus-worker", "confidence": 0.95}

Swarm Topology

Developed By

Atlas Institute — Code-Trainer / RTPI pipeline