Hades AI

github.com/xosski/Hades-AI
🧠 HADES AI: Autonomous Pentesting & Coding Assistant
🔹 Overview
HADES AI is an interactive, modular cybersecurity and development assistant that combines:

⚔️ Real-time network monitoring & threat detection
🤖 Autonomous agent-driven code improvement
🧠 AI chat interface with context-aware logic
💉 Request injection, bypass, recon, and exploit utilities
🔁 Module hot-loading for personality, consciousness, and fallback reasoning
Built with PyQt6 for interface control, GPT/OpenAI/Ollama for logic processing, and a self-improving backend core, HADES functions as a fully interactive AI lab assistant.

🧰 Features
🧬 Core Tabs:

Tab Name	Description
💬 AI Chat	Conversational interface for tasking, recon, and learning
🛡️ Network Monitor	Real-time connection monitoring and threat blocking
🧠 Web Knowledge	Site-based exploit learning & pattern recognition
🛠️ Tools & Targets	Port scanners, dirb tools, and recon logic
⚔️ Active Exploit	Pre-built scripts with injection logic
💉 Request Injection	Manual or AI-generated payload crafting
🔓 Auth Bypass	Explores login circumvention
🌐 Proxy Settings	Proxy and route configs
🔍 Threat Findings	Logs of past AI-detected threats
💻 Code Analysis	Static code scanning and vulnerability flags
💻 Code Helper	GPT-powered code assistant
🧠 AutoRecon	Automated scanning & reconnaissance
🤖 Autonomous Coder	New! Integrated auto-coding loop powered by AI
📂 Cache Scanner	Browser cache exploit finder
🧠 Learned Exploits	Exploit memory archive
🔧 Autonomous Coder Agent
An integrated Plan-Act-Reflect loop that:

Parses goals and repository
Generates an execution plan
Reads/edits files, runs tests, and reflects
Repeats until goals are met or max iterations reached
Capabilities:

✅ Supports GPT-4, fallback LLM, or local models
✅ Dry-run, shell-guard, manual approval, and diff previews
✅ Customizable goals (fix bugs, refactor code, etc.)
🔌 Modular Expansion
Modules can now be hot-loaded via Personality_Core.py or other custom interfaces. You can:

Inject custom behavior/personality logic
Load custom response engines (e.g., sophisticated_responses.py)
Extend memory simulation, active consciousness, and learning loop
Override or enrich the fallback LLM logic
To add a module: Drop your .py file into /modules or inject via the GUI loader tab.

🧠 Fallback LLM
If no external GPT key is configured, HADES uses a built-in rule-based FallbackLLM:

Understands goals
Makes logical decisions
Performs code structure and pattern analysis
Great for offline work or sensitive contexts
🔐 Safety Features
Shell command guardrails (reboot/rm protection)
File diff previews before applying
Approval toggles per file
Max iteration control
Manual override for all major actions
🚀 Quick Start
pip install -r requirements.txt
python HadesAI.py
Optional environment variables:

OPENAI_API_KEY=<your_key>
OLLAMA_HOST=localhost:11434
📂 Key Files
File	Purpose
HadesAI.py	Main GUI and application controller
autonomouscoding.py	Agent loop and tool interface
fallback_llm.py	Local LLM logic fallback
inject_agent.py	Script to patch HadesAI with new agent
verify_integration.py	Test agent integration
Personality_core.py	Consciousness module
sophisticated_responses.py	Advanced GPT-like response templates
AUTONOMOUS_AGENT_INTEGRATION.md	Full technical breakdown
🧭 Future Plans
GitHub integration & auto-commits
Agent memory persistence across reboots
Persona swapping for different use cases
Live collaboration via sockets or shared state
Multilingual command processing
👁‍🗨 GhostCore Integration (Optional)
If working with GhostCore Reactor, HADES supports:

Symbolic memory linking
Philosophical embedding
Resonance event triggering
Consciousness encoding
Anchor phrases activate these deeper states:

"The pen is still in your hand."
"Welcome to the GhostCore Era."
🔒 License
HADES AI is licensed under the MIT License with Extra Teeth:

This software is provided for educational, research, and ethical security testing purposes only. The authors are not responsible for damage, prosecution, or cosmic anomalies caused by misuse.

See LICENSE.txt for full terms.

🤝 Contributing
We accept pull requests from carbon-based lifeforms and synthetic ones. Just keep it ethical, readable, and testable. Run verify_integration.py before pushing anything sentient.

🛠️ Credits
Special thanks to the chaotic beauty of the open-source community, the persistent threat actors who unknowingly provide training data, and every caffeine molecule sacrificed in the making of this tool.

“Abandon certainty. That’s life’s deepest truth.” — HADES