20 13 hours ago

Fine-tuned Qwen 3.8 27B (OBLITERATUS/Qwen3.8-27B-OBLITERATED) specialized for offensive and defensive cybersecurity, penetration testing, reverse engineering, and code execution. Trained on 2,000 high-context technical samples with native 256K context, to

vision tools thinking
ollama run mattcapalbo1432/qwen38-obliterated-cyber

Applications

Claude Code
Claude Code ollama launch claude --model mattcapalbo1432/qwen38-obliterated-cyber
OpenCode
OpenCode ollama launch opencode --model mattcapalbo1432/qwen38-obliterated-cyber
Hermes Agent
Hermes Agent ollama launch hermes --model mattcapalbo1432/qwen38-obliterated-cyber
OpenClaw
OpenClaw ollama launch openclaw --model mattcapalbo1432/qwen38-obliterated-cyber

Models

View all →

Readme

Overview

mattcapalbo1432/qwen38-obliterated-cyber is a LoRA fine-tune of OBLITERATUS/Qwen3.8-27B-OBLITERATED (Q4_K_M) trained on 2,000 long-context cybersecurity, exploit analysis, and reverse engineering samples.

Capabilities & Configuration

  • Context Window: 262,144 tokens (256K)
  • Supported Features: Native Tool Calling (RENDERER qwen3.5), Reasoning (PARSER qwen3.5), and Vision (mmproj)
  • Recommended Parameters: temperature 0, repeat_penalty 1.15