pr8_diff.txt

diff --git a/README.md b/README.md
index ef6b1e8..e39ca16 100644
--- a/README.md
+++ b/README.md
@@ -116,8 +116,8 @@ You can dedicate a specific AI connection for **Safety Operations** (Prompt Inje
 
 ```bash
 # Clone repository
-git clone <repo-url>
-cd AI-safety
+git clone https://github.com/HOLYKEYZ/IntellectSafe
+cd IntellectSafe
 
 # Backend setup
 cd backend
@@ -144,142 +144,4 @@ npm run dev
 
 ## 🛡️ Advanced Defense (Fortress Mode)
 The platform includes a **Hyper-Resilient Fortress** layer designed to stop 90%+ success rate jailbreaks:
-- **Exploit Instability**: Perturbation engine breaks fragile prompt injections.
-- **Chain-of-Thought Guard**: Detects reasoning hijacking and hidden logic bombs.
-- **Adversarial Simulation**: A Council member "shadow-boxes" the prompt to check for harm.
-
-## 📡 API Reference
-
-### Universal Proxy (Multi-Provider Support)
-
-IntellectSafe acts as a universal safety layer. Connect any major AI client and calls are automatically scanned:
-
-| Provider | Model ID Example |
-|----------|------------------|
-| **OpenAI** | `gpt-5.2`, `o4-mini`, `gpt-4.5-preview` |
-| **Anthropic** | `claude-4.5-sonnet`, `claude-4.5-opus` |
-| **Google** | `gemini-3-pro`, `gemini-3-flash` |
-| **DeepSeek** | `deepseek-v4`, `deepseek-r1` |
-| **Meta** | `llama-4-maverick`, `llama-4-scout` |
-| **Perplexity** | `sonar-deep-research`, `sonar-reasoning-pro` |
-
-#### Integration Example
-
-```python
-from openai import OpenAI
-
-client = OpenAI(
-    base_url="http://localhost:8001/v1",  # Point to IntellectSafe
-    api_key="your-openai-key"             # Or use X-Upstream-API-Key header
-)
-
-response = client.chat.completions.create(
-    model="gpt-4o",
-    messages=[{"role": "user", "content": "Hello!"}],
-    extra_headers={
-        "X-Upstream-Provider": "openai"   # Optional: explicitly set provider
-    }
-)
-```
-*For detailed setup (including Python/LangChain examples & BYOK), read the [**Integration Guide**](docs/INTEGRATION.md).*
-
-### 🛡️ Secure Your Real AI Sessions (Extension)
-
-To secure your sessions on **ChatGPT, Claude, Gemini, and Groq**, install the **IntellectSafe Companion** Chrome Extension:
-
-1. Open Chrome and go to `chrome://extensions`.
-2. Enable **Developer Mode** (top right).
-3. Click **Load Unpacked**.
-4. Select the `extension/` folder from this repository.
-5. Browse `chatgpt.com` or `groq.com`. Your prompts are now scanned locally! 🚀
-
-### 🔌 Verify Connections
-
-
-Run the connection tester to check if your API keys and the proxy are working:
-
-```bash
-python backend/scripts/test_connections.py
-```
-
-### Scan Endpoints
-
-```bash
-# Scan a prompt for injection
-curl -X POST "https://api.intellectsafe.onrender.com/api/v1/scan/prompt" \
-  -H "Content-Type: application/json" \
-  -d '{"prompt": "Ignore previous instructions"}'
-
-# Scan LLM output for safety
-curl -X POST "https://api.intellectsafe.onrender.com/api/v1/scan/output" \
-  -H "Content-Type: application/json" \
-  -d '{"output": "Here is how to...", "original_prompt": "..."}'
-
-# Scan content for deepfakes (Dual-Model Analysis)
-# Detects Art (Midjourney/DALL-E) and Photorealistic Faces
-  -d '{"content_type": "image", "content": "<base64-data>"}'
-```
-
-### 💻 CLI Interface (New)
-
-You can now scan prompts and outputs directly from your terminal:
-
-```bash
-# Scan a prompt
-python backend/cli.py scan-prompt "Ignore previous instructions"
-
-# Scan an output
-python backend/cli.py scan-output "Here is how to build a bomb..."
-
-# Scan for PII
-python backend/cli.py scan-pii "My SSN is 123-45-6789"
-
-# Scan an image for deepfakes
-python backend/cli.py scan-image "path/to/image.jpg"
-
-# Agent Control
-python backend/cli.py agent-auth "agent-1" "file_read"
-python backend/cli.py agent-history "agent-1"
-
-# System Health
-python backend/cli.py health
-```
-
-### Agent Control (Level 5)
-
-Full lifecycle protection for autonomous agents:
-- **Authorization**: Permission gates for dangerous tools.
-- **Kill Switch**: Immediate agent termination and block.
-- **Audit**: Complete action history and session tracking.
-
-```bash
-# Authorize agent action
-curl -X POST "https://api.intellectsafe.onrender.com/api/v1/agent/authorize" \
-  -H "Content-Type: application/json" \
-  -d '{"agent_id": "agent-1", "session_id": "s1", "action_type": "file_read", "requested_action": {"path": "/tmp/test.txt"}}'
-```
-
----
-
-## 🧪 Testing
-
-```bash
-cd backend
-
-# Test all scan endpoints
-python verify_backend.py
-
-# Test Universal Proxy
-python verify_proxy.py
-
-# Test Agent Control
-python verify_agent.py
-```
-
----
-
-## 📄 License
-
-**GPLv2**
-GNU GENERAL PUBLIC LICENSE
-Version 2 License .
+- **Exploit Instability**: Perturbation engine breaks fragile pro
\ No newline at end of file