Eiseron AI Agent Guidelines

This repository contains the official engineering constitution for Eiseron. It defines the standards, workflows, and skills required for both human developers and AI agents (Cursor, Windsurf, etc.) to maintain a high-quality, secure, and maintainable codebase.

Core Constitution

• 00-ai-conventions.md: Fundamental rules for AI interaction, communication, and basic git etiquette.

Engineering Skills

The following modules provide detailed standards for various engineering domains:

1. Architecture & Design
   • 01-architecture.md: Clean Code, Phoenix/Vertical Dev, and Modular Monolith principles.
2. Data & Storage
   • 02-database-standards.md: Database naming, schema, and auditing rules.
3. Security & Privacy
   • 03-security.md: Authentication, authorization, and secure-by-default standards.
4. Frontend & UX
   • 04-frontend.md: Pure CSS, Semantic HTML, and Component-Driven Development.
5. Dependency Management
   • 05-dependency-management.md: Rules for selection, justification, and security of third-party libraries.
6. Quality Assurance
   • 06-testing-bdd.md: Testing Pyramid, BDD, and single-assertion patterns.
7. Infrastructure & DevOps
   • 07-docker-infra.md: Standardized Docker workflows and environment isolation.
8. Observability & Reliability
   • 08-observability-ops.md: Logging, monitoring, and health standards.
9. Workflow & Version Control
   • 09-git-workflow.md: Advanced Git strategies, release cycles, and branch protection.
10. Communication & Knowledge
    • 10-documentation.md: Documentation standards (Sync, Format, and Types).

Quick Installation (Global SSoT)

To enforce these global standards on your local machine and ensure AI agents always have the correct context, run the bootstrap installer:

curl -fsSL https://raw.githubusercontent.com/eiseron/eiseron-agents/main/install.sh | bash

This script is idempotent. It securely clones the repository to ~/.eiseron-agents, sets up a background auto-update task, and injects the global pointers into your AI IDE configurations (Cursor, Windsurf).

Usage

AI Agents must perform "Self-Discovery" at the start of every session by:

1. Listing the files in this repository to identify available skills.
2. Reading 00-ai-conventions.md and AGENTS.md.
3. Reading specific skill modules on-demand (lazy-loading) as required by the current objective to preserve context tokens.

All code generated must strictly adhere to these guidelines. Any deviation requires explicit user approval.