☠️ CyberHacker Suite

CyberHacker is a custom AI-native extension suite that integrates directly into any existing Kali Linux environment. It is not a custom OS image or a standalone distribution. Instead, it is a deeply integrated platform where Large Language Models (LLMs) and agentic AI pipelines act as first-class citizens alongside your existing Kali tools.

LLMs and vision-language models are integrated into the Kali Linux core and exposed as native OS services. Every AI agent that has access to OS-level capabilities operates under strict, non-negotiable guardrails — preventing unauthorized actions, data exfiltration, and destructive commands unless explicitly permitted in a controlled, authorized lab environment.

Designed for: Authorized cybersecurity research, penetration testing labs, CTF practice, defensive tooling, and AI-augmented red/blue team operations.

CyberHacker embeds AI into your system through the following stack:

All models run 100% locally — zero cloud APIs required. Optimized for an NVIDIA RTX 5060 (8 GB VRAM) with phase-scheduled GPU loading to avoid OOM errors.

The CyberHacker suite is designed to be more than just a tool—it acts as a personalized cybersecurity companion.

• Reinforcement Learning (RL) & Adaptation: The system actively uses RL to learn from your interactions, adapt to your specific preferences, and grow alongside your workflows. It builds an understanding of your methodology over time.
• Persistent Identity: Users can associate with the suite as a consistent companion. Your preferences, historical context, and customized workflows are retained as part of the companion's core memory.
• Decoupled Intelligence: The "brain" is interchangeable. When you swap the underlying LLM (whether switching to a different local raw model or connecting to an external API), you are simply upgrading the companion's raw intelligence and extent of capabilities. The companion's identity and understanding of you remain intact, allowing it to leverage the new model's power without losing its personalized context.

Every AI agent in the CyberHacker suite operates with strictly enforced guardrails baked into the service layer:

• Authorization checks — Agents cannot execute against targets outside an explicitly declared scope.
• Non-destructive defaults — Destructive actions (file deletion, payload deployment, privilege escalation) require explicit human confirmation even with root access.
• Audit logging — Every agentic action is logged to a tamper-evident local audit trail.
• Scope-limited tool calls — LangGraph pipelines run inside sandboxed execution contexts that enforce network, filesystem, and process constraints.
• Ethical use enforcement — The system is architecturally designed for authorized environments: CTF labs, personal test machines, and systems the user has legal permission to operate against.

The AI intelligence powering the CyberHacker suite is trained on knowledge produced by the CyberX Dataset Factory — a fully local, automated pipeline that ingests cybersecurity textbooks and produces structured training data.

graph TD;
    A[📚 Cybersecurity Books / PDFs] -->|Render Pages| B[Universal Page Renderer];
    B --> C[👁️ Visual Path: ColQwen2 Embeddings];
    B --> D[📝 Text Path: OCR / Docling / Layout Detection];
    C --> E[Master Knowledge Corpus];
    D --> E;
    E --> F[Knowledge Atoms: Techniques, Tools, Vulnerabilities];
    F --> G[Scenario & Decision Generation via Qwen2.5];
    G --> H[Verification + Visual Grounding Check];
    H --> I[(Canonical Dataset)];
    I --> J[Qwen2.5 Fine-Tune Adapter];
    I --> K[Future LLM Adapter];
    J --> L[🤖 CyberHacker AI Core];

Unlike standard RAG pipelines that flatten PDFs into linear text, the factory treats every page as a visual object first. This preserves the spatial relationships between diagrams, tables, callouts, and code blocks that plain text extraction destroys.

The CyberHacker suite integrates seamlessly with your Kali Linux base:

• Native agentic services — LangGraph pipelines registered as systemd services, available from the first shell session.
• Workflow Enhancements — Integrates with your desktop environment for a focused security-research workflow.
• Tool Interoperability — Works alongside and augments all standard Kali tools out-of-the-box.

• 📋 PRD — Full product requirements for the dataset factory & OS integration.
• 🛠️ Tech Stack — Models, databases, and pipelines in detail.
• ⚠️ Constraints & Design Rules — VRAM limits, privacy mandates, and architectural guardrails.

All commits are gated by a strict pre-push pipeline:

• Ruff — Python linting & formatting
• Prettier / ESLint — Web asset formatting
• PMD — Java anti-bloat checks
• repo-secret-scan — Full repo secret detection before every push

Run setup-git2.ps1 to initialize local Git hooks after cloning.