Genesis Platform

Welcome to the Genesis Platform, a comprehensive, self-contained ecosystem for managing a secure, multi-tenant, and AI-driven software development lifecycle. This monorepo contains all the services required to operate the platform, from the root Certificate Authority to the autonomous agents that perform development tasks.

Core Philosophy: The Phoenix Protocol

The platform is designed around a core philosophy called The Phoenix Protocol, which applies the principles of safety-critical systems engineering to the domain of autonomous AI development. This means every task undertaken by an agent is not just executed, but is also automatically tested, reviewed, and verified through a multi-stage pipeline to ensure the highest degree of quality and reliability.

Our key principles are:

• Automation First: All core operations are driven by APIs, enabling a fully autonomous development process orchestrated by AI agents.
• Security by Default: The platform uses a robust, certificate-based authentication system to ensure that all interactions are secure.
• Infrastructure as Code: The entire platform is defined as code, from the Dockerfiles to the setup scripts, making it reproducible, auditable, and manageable.
• Verifiable Autonomy: Agent-generated solutions are not blindly trusted; they are rigorously validated at each step of the development lifecycle.

Services

• ca: The root of trust. Issues short-lived SSH certificates to agents.
• manager: The central control plane and build orchestrator. Provides APIs to manage repositories, agents, missions, and just-in-time agent environments.
• cortex: The AI "brain." A lightweight gateway to powerful LLMs via the OpenRouter API.
• backup: The archival service. Performs encrypted, stateful backups of the entire platform.
• recovery: The restoration service. Enables full or granular recovery from backups.
• security: The immune system. Provides automated detection and response for security incidents.

Getting Started

To get started, you will need Docker and Docker Compose. Run the following setup scripts in order:

1. ./ca/setup_ca.sh
2. ./cortex/setup_cortex.sh
3. ./manager/setup_manager.sh
4. ./backup/setup_backup.sh
5. ./recovery/setup_recovery.sh
6. ./security/setup_security.sh

Once complete, the core services will be running. You can verify their status:

• Manager API: http://localhost:8001/docs
• Cortex API: http://localhost:8002/docs

Agent & Mission Lifecycle

The platform operates on a Hierarchical Hybrid Planner model, orchestrated by the Phoenix Protocol. This combines strategic planning with a rigorous, multi-stage verification process and a dynamic build system.

1. Implementation (Planner -> Worker): A Planner agent creates an implementation plan. A Worker agent executes it in a custom-provisioned environment, committing the feature code.

2. Verification (Auditor -> Tester): An Auditor agent creates a verification plan (static analysis, E2E test generation). A Tester agent executes this plan in a separate, custom-provisioned test environment, committing the new E2E test suite back to the repository.

3. Review (Reviewer Agent): A Reviewer agent performs a final, qualitative "peer review" of the code, assessing it for maintainability, security, and elegance.

4. Completion: Only after passing all three stages is the mission marked as FULLY_VERIFIED, leaving behind not only a working feature but also a new set of tests and a qualitative review record.

This architecture is detailed further in phoenix_protocol.md and agent_planner_architecture.md.