# .Agent Rules - Agent Scaffolding

## Module Overview

**Purpose**: This directory serves as the immutable foundational architecture and behavioral constitution for all AI agents (including Antigravity, Jules, and custom MCP clients) operating within the GNN ecosystem.

**Pipeline Step**: Global Configuration / Cognitive Scaffolding

**Category**: Governance / Policy

**Status**: Maintained

**Version**: 1.6.0

---

## Core Functionality

### Primary Responsibilities

1. **Policy Enforcement**: Defines non-negotiable architectural mandates (e.g., real-implementation testing, Thin Orchestrator Pattern, and God-Class Decomposition routines).
2. **Behavioral Framing**: Governs how agents interact with the file system, manage technical debt, and format outputs.
3. **Quality Assurance**: Outlines testing thresholds, linting guidelines, and error-handling paradigms directly accessible via the MCP framework.

### Available Documentation Rulesets

- **`architecture.md`**: Global systems architecture patterns.
- **`testing.md` / `quality.md`**: Testing frameworks and static analysis mandates.
- **`dependencies.md` / `render_frameworks.md`**: Technical limitations and optimal libraries.
- **`gnn_standards.md`**: Domain-specific logic constraints for Active Inference modeling.

## Implementation Details

These rules are parsed natively by the local IDE or the Personal AI Infrastructure (PAI) orchestrator before execution initialization. No Python objects are executed here; it acts strictly as RAG context and persistent behavioral overrides.