The world’s most comprehensive digital armor, now with developer-first extensibility.
Build, connect, and deploy next‑gen security agents—without subscription chains.

Imagine a transparent, adaptable security fabric that wraps around every layer of your digital life—from kernel-level threat interception to cloud-based AI triage. That’s what we’ve rebuilt.

This repository contains the open-architecture blueprint for McAfee LiveSafe 16.0 R50, a security platform designed to be:

• Modular: Swap out detection engines (signature, heuristic, behavioral) without rebooting.
• API‑native: Integrate with your existing SIEM, SOAR, or custom Python/Go scripts in minutes.
• Context‑aware: Uses local LLM reasoning (via OpenAI/Claude) to decide why a file is suspicious, not just if.
• Zero‑obligation: No licenses. No subscriptions. Use the core engine forever.

A good antivirus catches viruses. A great one catches intent.
This suite learns from 40+ years of threat intelligence and gives you the keys to the engine room.

graph TD
    A[File/Process Trigger] --> B{Signature Match?}
    B -->|No| C{Heuristic Analysis}
    B -->|Yes| D[Auto-Quarantine + Log]
    C -->|Suspicious| E[Behavioral Sandbox]
    C -->|Clean| F[Allow + Cache]
    E --> G[LLM Reasoning Layer<br>OpenAI / Claude]
    G --> H{Malicious Intent?}
    H -->|Yes| I[Deep Remediation]
    H -->|Maybe| J[User Alert + Context]
    H -->|No| K[Threat Intelligence Feed Update]
    I --> L[Restore System State]
    J --> M[Interactive Console]
    K --> N[Community Blacklist]

Legend:

• 🛡️ Signature Match – fast, deterministic scanning (MD5/SHA256 libraries)
• 🧠 Heuristic Analysis – behavioral pattern recognition (memory, registry, network)
• 🤖 LLM Reasoning Layer – your own OpenAI or Claude API key interprets ambiguous behavior
• 🔁 Threat Intelligence Feed – anonymized data points improve detection for everyone

Define how LiveSafe behaves per environment using a YAML profile:

profile:
  name: "Workstation-HighSecurity"
  scan:
    realtime: true
    scheduled: "weekly"
    deepScan: true  # scans inside archives + memory
  decisionEngine:
    useOpenAI: true
    openaiModel: "gpt-4o-mini"
    useClaude: true
    claudeModel: "claude-3-haiku-20240307"
    fallbackMode: "aggressive"  # quarantine on any LLM uncertainty
  notifications:
    severityFilter: "medium"  # only alert on medium threats+
    pushEndpoint: "https://my.siem.internal/alerts"
  sandbox:
    timeout: 60 # seconds
    snapshot: true  # save system state before test

Run a one‑time deep scan on a directory with custom API keys:

livesafe-console --scan /home/user/downloads \
                 --profile enterprise_default \
                 --openai-key sk-your-key-here \
                 --claude-key sk-ant-your-key-here \
                 --log-level verbose \
                 --output-format json

Expected output snippet (JSON):

{
  "scannedFiles": 2471,
  "threatsFound": 3,
  "actionTaken": ["quarantined", "remediated"],
  "llmReasoning": "File 'invoice_2026.docm' executes PowerShell within 2 seconds. High risk."
}

Looking for a McAfee LiveSafe 16.0 R50 open-source security suite?
Need a lightweight antivirus with AI reasoning for developers?
This project is built for enterprise threat detection without vendor lock-in.

• Alternative to subscription‑based security – run your own detection logic.
• Integrates with OpenAI GPT‑4o and Claude 3.5 – bring your own key.
• Cross‑platform malware analysis – Windows, macOS, Linux, Android, iOS.
• 2026‑ready – designed for the next‑gen threat landscape.

Bring your own OpenAI or Claude key to augment decision‑making:

# In your profile.yaml:
reasoning:
  provider: "claude"
  apiKey: "sk-ant-your-key-here"
  model: "claude-sonnet-4-20260514"
  instruction: "Explain in plain English why this binary is suspicious."

Benefits of LLM integration:

• 🧠 Explainable AI – every quarantine decision includes a plain‑English reason.
• 📚 Zero‑day detection – LLM spots novel patterns by comparing to its training knowledge.
• 🚦 Adaptive urgency – low confidence? Escalate to human analyst.
• 🔐 Local‑only mode – your API calls stay encrypted; we never log raw keys.

Built with Flutter + WebAssembly. On a 27” monitor you get a full threat map; on a phone you get a clean status bar. Transitions feel like glass – no lag, no page reloads.

We use ICU MessageFormat for dynamic copying. Threat names appear in your local dialect (e.g., “Trojan banker” becomes “Trojan bancaire” in French). The interface responds to your OS locale automatically.

• Chatbot: Responds in < 500ms for common questions.
• Human Escalation: Critical incidents get a real security engineer in < 3 minutes.
• Video Walkthroughs: Pre‑recorded guides for complex features.

This project is provided “as is” for educational and research purposes.

• The authors are not liable for any unintended consequences of using this software.
• You are responsible for complying with all applicable laws in your jurisdiction regarding security tools.
• This repository does not encourage bypassing any security measures or installing unauthorized software.
• Use of OpenAI, Claude, or any third‑party API is subject to their respective terms of service.
• No warranty is offered – test thoroughly in a sandboxed environment before production use.

Security is a responsibility, not a product.
Use this tool to defend, not to exploit.

This project is distributed under the MIT License.
You are free to use, modify, and distribute the software, provided the original copyright notice is included.

McAfee LiveSafe 16.0 R50 – The security engine that learns like you do.
Build your defense. Own your data. Connect with the community.

Not a crack. Not a crack. Not a patch. A genuine, open‑source security framework for the bold.