Eve library

Psyche Security

Other

Revision 1

On this page

Comprehensive security and privacy framework for the Psyche AI Virtual Assistant Platform.

Part of the Oshun Platform.

Overview#

This library provides enterprise-grade security capabilities for AI virtual assistants, including authentication, authorization, data protection, AI-specific security, privacy compliance, and security monitoring.

Modules#

Authentication & Authorization#

  • OAuth2 / JWT token management
  • Multi-factor authentication (MFA)
  • Role-based access control (RBAC)
  • Session management
  • API key management
  • SSO integration

Data Protection#

  • AES-256 encryption at rest
  • TLS 1.3 for data in transit
  • PII detection and classification
  • Data retention policies
  • Secure key management

AI Security#

  • Prompt injection detection
  • Output filtering and sanitization
  • Hallucination detection
  • Bias detection and mitigation
  • Model access control

Privacy Controls#

  • Consent management (GDPR)
  • Data subject rights handling
  • Anonymization and pseudonymization
  • Privacy preference management
  • Data minimization

Compliance#

  • HIPAA compliance manager
  • SOC 2 controls
  • Audit logging
  • Compliance reporting
  • Policy enforcement

Security Monitoring#

  • Intrusion detection (IDS)
  • Anomaly detection
  • Rate limiting
  • Threat intelligence
  • Alert management

Project Structure#

text
src/security/
├── __init__.py           # Package exports
├── auth/                 # Authentication & Authorization
│   ├── types.py          # Auth data types
│   ├── authentication.py # Auth service
│   └── authorization.py  # RBAC/permissions
├── data_protection/      # Encryption & Data Safety
│   ├── encryption.py     # Crypto services
│   ├── pii_detector.py   # PII detection
│   └── classifier.py     # Data classification
├── ai_security/          # AI-Specific Security
│   ├── prompt_injection.py
│   ├── output_filter.py
│   └── hallucination.py
├── privacy/              # Privacy Controls
│   ├── consent.py        # Consent management
│   ├── dsr.py            # Data subject rights
│   └── anonymizer.py     # Anonymization
├── compliance/           # Regulatory Compliance
│   ├── hipaa.py          # HIPAA compliance
│   ├── audit.py          # Audit logging
│   └── reporter.py       # Compliance reports
└── monitoring/           # Security Monitoring
    ├── intrusion.py      # IDS
    ├── anomaly.py        # Anomaly detection
    └── alerts.py         # Alert management

Development#

Using Nx#

bash
# Install dependencies
nx install psyche-security

# Build
nx build psyche-security

# Linting
nx lint psyche-security
nx lint-fix psyche-security

# Formatting
nx format psyche-security
nx format-check psyche-security

# Type checking
nx typecheck psyche-security

# Testing
nx test psyche-security
nx test-cov psyche-security

Direct Commands#

bash
cd apps/psyche/security

# Install dependencies
poetry install

# Run tests
poetry run pytest

# Lint
poetry run ruff check src tests

Usage Examples#

Authentication#

python
from security import get_authentication_service

auth = get_authentication_service("your-secret-key")

# Register a user
user = await auth.register_user("user@example.com", "password")

# Authenticate
token = await auth.authenticate("user@example.com", "password")

# Verify token
claims = await auth.verify_token(token.access_token)

Authorization#

python
from security import get_authorization_service

authz = get_authorization_service()

# Check permission
if authz.has_permission(user.user_id, "resource:read"):
    # Access granted
    pass

# Check role
if authz.has_role(user.user_id, "admin"):
    # Admin access
    pass

AI Security#

python
from security import get_ai_security_service

ai_security = get_ai_security_service()

# Check for prompt injection
result = await ai_security.check_prompt(user_input)
if result.is_safe:
    # Process the prompt
    pass
else:
    # Handle potential attack
    pass

# Filter AI output
safe_output = await ai_security.filter_output(model_response)

Privacy#

python
from security import get_consent_manager

consent = get_consent_manager()

# Record consent
await consent.record_consent(
    user_id=user.user_id,
    purpose="analytics",
    granted=True,
)

# Check consent
if await consent.has_consent(user.user_id, "analytics"):
    # Collect analytics
    pass

Configuration#

Environment Variables#

Variable Description
SECURITY_SECRET_KEY Master secret for JWT signing
ENCRYPTION_KEY Key for data encryption
AUDIT_LOG_LEVEL Audit logging verbosity
RATE_LIMIT_REQUESTS Rate limit per window
RATE_LIMIT_WINDOW Rate limit window in seconds

License#

Proprietary - Oshun Platform