🧬 What is Aether-CyberAgent?

Aether-CyberAgent is an elite, autonomous, multi-agent AI security platform and terminal co-pilot. It fuses deterministic computer science (AST parsing, graph theory, formal logic) with probabilistic AI (Google Gemini function calling) to create a self-healing, autonomic software engine.

Version 0.2.0 introduces the Antigravity Interactive Agent REPL, transforming Aether from a CLI scanner into a full autonomous software developer and security team co-pilot that can converse, write scripts, execute terminal commands, and self-correct when errors occur.

✨ Key Features (v0.2.0)

• 💬 Interactive Agent REPL — Converse with Aether dynamically in a rich terminal interface.
• 🛠️ Autonomous Tool Execution — Aether can read/write files and execute shell scripts safely.
• 🔄 Self-Correction Loop — If a script fails, the Gold Team intercepts the stderr and feeds it back to the Yellow Team for autonomous auto-fixing (up to 3 retries).
• 🔍 Deterministic Vulnerability Detection — Python AST-based scanning, not regex guessing.
• 🧠 AI-Powered Remediation — Google Gemini generates type-annotated, secure code patches.
• 📐 Mathematical Verification — Z3 SMT Solver proves correctness, Hypothesis fuzzes edge cases.
• 📊 Live Dashboard — Real-time Streamlit visualization of the security pipeline.
• 📋 SARIF v2.1.0 Reports — Native GitHub Security tab integration.

🏛️ Architecture

Aether-CyberAgent implements a defensive Cybersecurity Team Spectrum model. Each agent maps to a distinct security role — there are no offensive components. This is a pure defensive, self-correcting system.

The Team Spectrum

┌──────────────────────────────────────────────────────────────────────┐
│                     🥇 GOLD TEAM (Orchestrator)                      │
│              Autonomic Self-Healing Loop · Max 3 Retries             │
│                                                                      │
│  ┌────────────┐   ┌────────────┐   ┌────────────┐   ┌────────────┐  │
│  │ 🔵 BLUE    │──▶│ 🟡 YELLOW  │──▶│ 🟣 PURPLE  │──▶│ 🟢 GREEN   │  │
│  │ AST Audit  │   │ AI Patch   │   │ Z3 Verify  │   │ Tool Run   │  │
│  │ + GraphRAG │   │ + Scripts  │   │ + Hypothes │   │ + Sandbox  │  │
│  └────────────┘   └─────▲──────┘   └─────┬──────┘   └─────┬──────┘  │
│                         │                 │                 │         │
│                         │    ❌ FAIL      │                 │         │
│                         └─────────────────┘                 │         │
│                                                             ▼         │
│                                                     ┌────────────┐   │
│                                                     │ ⚪ WHITE   │   │
│                                                     │ SARIF v2.1 │   │
│                                                     │ + Reports  │   │
│                                                     └────────────┘   │
└──────────────────────────────────────────────────────────────────────┘

🚀 Getting Started

Prerequisites

• Python 3.11+
• Docker (running daemon, for sandbox execution)
• Google Gemini API Key (Get one free)

Installation

# Clone the repository
git clone https://github.com/mrx231066/Aether-CyberAgent.git
cd Aether-CyberAgent

# Install the CLI package
pip install -e .

💡 Usage

Launch Interactive Agent REPL (New in v0.2.0)

Simply type aether to launch the Antigravity Interactive Agent REPL. It will dynamically discover available Gemini models and prompt for your API key if not configured.

aether

Supported Slash Commands:

• /help — Displays interactive command matrix & agent team status.
• /scan [path] — Executes full background multi-agent security scan.
• /model — Opens interactive model switcher menu.
• /auth — Updates stored API key/config.
• /status — Displays dependency graph metrics & memory state.
• /run <script> — Safely runs a script in the execution sandbox.
• /clear — Clears terminal output.
• /exit or /quit — Closes REPL session.

Autonomous Script Execution & Self-Correction: If you ask Aether to write a script in the chat, it will generate it, write it to disk, and execute it. If it fails, the Gold Team intercepts the traceback and auto-fixes the code up to 3 times!

Run a Standalone Security Scan

# Scan the current directory
aether scan .

# Scan a specific path
aether scan ./src/my_project

Launch the Visual Dashboard

# Start the Streamlit real-time dashboard
aether dashboard

Verify a Specific File

# Run formal verification on a specific file patch
aether verify ./src/module.py

📁 Project Structure (v0.2.0)

aether-cyberagent/
├── aether/
│   ├── ai/
│   │   └── gemini_client.py       # Dynamic model discovery & Structured Output
│   ├── agents/
│   │   ├── blue_auditor.py        # Static AST parser
│   │   ├── yellow_patcher.py      # Refactoring & script generator agent
│   │   ├── purple_verifier.py     # Z3 formal logic & property testing
│   │   └── gold_autonomic.py      # Self-correction execution loop
│   ├── engine/
│   │   ├── graph_memory.py        # NetworkX repository dependency graph
│   │   ├── sandbox.py             # Docker runner
│   │   └── tools.py               # Autonomous tool execution engine (read/write/shell)
│   ├── reports/
│   │   └── sarif.py               # White team SARIF v2.1.0 exporter
│   ├── dashboard/
│   │   └── app.py                 # Streamlit visualizer
│   └── cli/
│       ├── main.py                # Typer CLI entrypoint
│       └── interactive.py         # Antigravity Interactive REPL & Slash Commands
├── tests/
│   ├── test_full_pipeline.py      # Scan pipeline tests
│   └── test_interactive_repl.py   # REPL & tool engine tests
├── pyproject.toml                 # Package configuration
└── README.md                      # Documentation

🤝 Contributing

Contributions are welcome! Please read our contributing guidelines before submitting PRs.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License — see the LICENSE file for details.