feat: add infrastructure security system (Layer 4 - Part 3/5)

Implements Agentic Infrastructure Layer 4 - Security Component
Part of v0.11.0 implementation roadmap

Added:
- RBAC (Role-Based Access Control): Fine-grained permissions
  - Permission enum with predefined permissions
  - Role system with permission groups
  - Role inheritance support
  - User-role assignment
  - Permission checking and enforcement
  - Default roles (admin, user, readonly, tool_executor)
  - Decorator for permission requirements
  - Comprehensive permission categories:
    * Tool permissions (execute, register, delete)
    * File permissions (read, write, delete, execute)
    * Network permissions (http, https, socket)
    * Environment permissions (read, write, shell)
    * Memory permissions (read, write, delete)
    * Agent permissions (create, delete, configure)
    * Admin permissions (all access)

- Sandbox: Isolated execution environment
  - File access control (allowed/blocked paths)
  - Network access control
  - Environment variable control
  - Subprocess execution control
  - Resource limits support
  - Context manager and decorator support
  - Safe execution of untrusted code
  - Default secure configurations
  - Predefined sandboxes (readonly, network, tool)

- AuditLogger: Comprehensive audit trail
  - Structured audit events with levels
  - Security event logging
  - Access attempt tracking
  - Tool execution logging
  - Configuration change logging
  - Error logging
  - Event querying and filtering
  - Security compliance reporting
  - JSON and text format support
  - File and console output

- SecretsManager: Secure secrets storage
  - Encrypted storage (with rotation support)
  - Secret expiration tracking
  - Access counting
  - Environment variable integration
  - Metadata support
  - Secret rotation capabilities
  - Load/save from encrypted file
  - API key hashing utilities
  - Statistics and reporting

- Demo: infrastructure_security_demo.py
  - 5 comprehensive demos showing:
    1. RBAC with roles and permissions
    2. Sandbox for isolated execution
    3. Audit logging for compliance
    4. Secrets management
    5. Integrated security (all patterns combined)

Directory Structure:
react_agent_framework/
├── infrastructure/
│   └── security/
│       ├── __init__.py
│       ├── permissions.py       (373 lines)
│       ├── sandbox.py           (336 lines)
│       ├── audit.py             (367 lines)
│       └── secrets.py           (342 lines)

Security Features:
- Production-ready access control
- Defense in depth (multiple layers)
- Audit trail for compliance
- Secure credential management
- Protection against common attacks

Next Steps (Fase 1 continuation):
- Part 4/5: Cost Control (budget, rate limiter, quota)
- Part 5/5: Human-in-the-Loop (approval, intervention, feedback)

Related to: Agentic AI Layer 4 implementation
Target version: v0.11.0
Progress: 60% of Fase 1 completed (3/5 components)

Author: marcosf63