README.md

metadata

title: Optimax-agent
emoji: 🛡️
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 6.3.0
app_file: app.py
pinned: false
license: apache-2.0

🛡️ Optimax Security Agent

AI-powered Salesforce security analysis API that receives metadata from Salesforce orgs and returns comprehensive security assessments.

Architecture

Salesforce Org → Sends Metadata → Optimax Agent API → Returns Analysis

Key Feature: This API processes ONLY METADATA - no credentials are ever received or stored.

Features

• 🤖 AI-Powered Analysis using Salesforce CodeGen 350M
• 🔍 Permission Vulnerability Detection
• 👤 Identity & Access Management Analysis
• 🔐 Sharing Model Security Assessment
• 📊 Risk Scoring & Prioritization
• 💡 Actionable Recommendations

API Endpoints

POST /api/analyze

Main analysis endpoint for Salesforce integration.

Request:

{
  "org_id": "00Dxx0000001234",
  "org_name": "Your Organization",
  "users": [...],
  "profiles": [...],
  "permission_sets": [...],
  "login_history": [...],
  "sharing_settings": {...}
}

Response:

{
  "success": true,
  "overall_risk_score": 45,
  "risk_level": "MEDIUM",
  "critical_findings": [...],
  "high_findings": [...],
  "ai_executive_summary": "...",
  "ai_recommendations": [...]
}

Usage

From Salesforce Apex:

// 1. Collect metadata
Map<String, Object> metadata = MetadataExtractor.collectOrgMetadata();

// 2. Call Optimax Agent
HttpRequest req = new HttpRequest();
req.setEndpoint('callout:Optimax_Agent/api/analyze');
req.setMethod('POST');
req.setHeader('Content-Type', 'application/json');
req.setBody(JSON.serialize(metadata));

Http http = new Http();
HttpResponse res = http.send(req);

// 3. Parse response
Map<String, Object> analysis = (Map<String, Object>)JSON.deserializeUntyped(res.getBody());

Using cURL:

curl -X POST https://m8077anya-vishwakarma-optimax-agent.hf.space/api/analyze \
  -H "Content-Type: application/json" \
  -d @metadata.json

Testing

Visit the app in your browser to use the interactive testing interface.

Security

• ✅ No credential processing
• ✅ Metadata-only analysis
• ✅ Stateless operation
• ✅ Private space (access controlled)

Technical Details

• Framework: Gradio + FastAPI
• AI Model: Salesforce CodeGen 350M (350 million parameters)
• Analysis: Hybrid (AI + rule-based)
• Deployment: Hugging Face Spaces

---

Version: 2.0.0
Last Updated: January 2026