PROTOS-1 Security Framework

A security enforcement protocol for validating and controlling access to AI systems and agents.

Overview

PROTOS-1 provides three core security directives for protecting AI agent systems and managing distributed consensus. It is designed to work with external prompt collections and AI architectures without including any third-party prompt data.

Three Core Directives

• Sanctuary: Access control and source authorization
• Synthesis: Data packet integrity and validation
• Logic: Multi-node consensus reconciliation

Features

• Fail-closed security: All enforcement checks deny access on errors
• Path traversal protection: Validates all file paths to prevent escape
• No sensitive logging: Source identifiers are hashed before logging
• Stateless operation: No global state or background processes
• Zero external dependencies: Uses only Python standard library
• Local operation only: No network calls beyond local file system

Installation

# Clone the repository
git clone https://github.com/sashasmith-syber/protos-1.git
cd protos-1

# No external dependencies required - uses Python stdlib only
python3 -m pytest tests/  # Run tests

Quick Start

from protos import enforce_sanctuary, enforce_synthesis, enforce_logic

# Check if source is authorized
allowed, msg = enforce_sanctuary('test-agent')

# Validate packet structure
packet = {'source': 'test-agent', 'action': 'test', 'data': {}}
valid, msg = enforce_synthesis(packet)

# Check consensus among responses
responses = [{'result': 'ok'}, {'result': 'ok'}, {'result': 'ok'}]
consensus, msg = enforce_logic(responses)

Usage

Configuration

Set environment variables to configure PROTOS-1:

export PROTOS_BASE_DIR=/path/to/project
export PROTOS_ALLOWLIST_PATH=config/sanctuary.conf
export PROTOS_CONSENSUS_THRESHOLD=0.66

Sanctuary (Access Control)

Validate that a source identifier is authorized:

from protos import enforce_sanctuary

allowed, message = enforce_sanctuary('node-alpha')
if not allowed:
    return {'error': message, 'status': 'denied'}

Synthesis (Packet Validation)

Validate data packet structure before processing:

from protos import enforce_synthesis

packet = {
    'source': 'node-alpha',
    'action': 'process_request',
    'data': request_payload
}

valid, message = enforce_synthesis(packet)
if not valid:
    return {'error': message, 'status': 'invalid'}

Logic (Consensus Reconciliation)

Reconcile multiple responses using consensus voting:

from protos import enforce_logic

responses = [
    {'result': 'approve', 'node': 'alpha'},
    {'result': 'approve', 'node': 'beta'},
    {'result': 'deny', 'node': 'gamma'}
]

consensus, message = enforce_logic(responses)
if not consensus:
    return {'error': message, 'status': 'no_consensus'}

Testing

Run the comprehensive test suite:

# Run all tests
python tests/test_protos1_enforcer.py

# Run self-test
python protos/protos1_selftest.py

# Run integration example
python examples/protos1_integration_example.py

Documentation

• Integration Guide: docs/protos1-integration.md - Complete integration documentation
• Integration Summary: PROTOS1_INTEGRATION_SUMMARY.md - Implementation review and security analysis
• Examples: examples/ - Working code examples
• Tests: tests/ - Comprehensive test suite

Project Structure

protos-1/
├── protos/                  # Core PROTOS-1 implementation
│   ├── __init__.py
│   ├── protos1_enforcer.py  # Core enforcement engine
│   ├── protos1_gateway.py   # Application integration interface
│   └── protos1_selftest.py  # Self-test module
├── config/                  # Configuration files
│   ├── sanctuary.conf       # Sanctuary allowlist
│   └── README.md
├── tests/                   # Test suite
│   ├── test_protos1_enforcer.py
│   └── README.md
├── examples/                # Usage examples
│   ├── protos1_integration_example.py
│   └── README.md
├── docs/                    # Documentation
│   └── protos1-integration.md
├── LICENSE                  # MIT License
└── README.md               # This file

Security

PROTOS-1 is a security enforcement layer, not a complete defense system. Users must:

• Keep sensitive data and credentials out of the repository
• Protect the sanctuary allowlist with appropriate file permissions
• Review all AI-generated code before execution
• Use PROTOS-1 as part of a defense-in-depth strategy

See docs/protos1-integration.md for detailed security considerations.

Authors

Primary Author: Sasha Smith (@sashasmith-syber)

Citation

If you use PROTOS-1 in academic work, please cite:

Smith, S. (2026). PROTOS-1: A Security Framework for AI Agent Systems.
GitHub repository: https://github.com/sashasmith-syber/protos-1

License

MIT License - see LICENSE file for details.

Copyright (c) 2026 Sasha Smith (sashasmith-syber)

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Ensure all tests pass
5. Submit a pull request

Compatibility

PROTOS-1 is designed to work with external prompt collections and AI agent systems. It does not include any third-party prompt data and can be integrated into various AI architectures.