Version: 1.1.0
Last updated: 2026-02-08

This repository is a quickstart template for AI-augmented / agentic workflows built on a doctrine stack: a five-layer governance system that separates values, models, instructions, procedures, and structure for predictable, inspectable agent behavior.

Five layers with clear precedence ensure consistent, predictable agent behavior:

Guidelines (values) ← Highest precedence ↓ Approaches (models) ↓ Directives (instructions) ← Select tactics ↓ Tactics (procedures) ← Execute work ↓ Templates (structure) ← Lowest precedence 

How it works:

• Guidelines define values and preferences (what matters most)
• Approaches provide mental models and philosophies (how to think)
• Directives give explicit instructions and constraints (what to do)
• Tactics execute work procedurally (step-by-step how)
• Templates provide output structure contracts (format)

Key Principles:

• Directives select which tactics to invoke
• Tactics execute the work procedurally
• Human retains approval authority ("Human in Charge")
• Modular loading minimizes token consumption

📖 Read: Doctrine Stack Documentation
📖 Browse: Tactics Index
📖 See: Agent Profiles
📖 Reference: Glossary

doctrine/ # Portable agentic framework (git subtree distributable) ├── DOCTRINE_STACK.md # Five-layer governance framework ├── GLOSSARY.md # Standardized terminology ├── guidelines/ # Values and preferences ├── approaches/ # Mental models and philosophies ├── directives/ # Modular operational instructions (20+) ├── tactics/ # Procedural execution guides (9 tactics) ├── templates/ # Output structure contracts └── agents/ # Specialized agent profiles (20 agents) docs/ # Project intent and architecture ├── VISION.md # Repository purpose ├── architecture/ # Strategic decisions │ ├── architectural_vision.md │ ├── adrs/ # Architecture Decision Records │ └── diagrams/ # PlantUML C4 diagrams └── templates/ # Additional output templates .doctrine-config/ # Project-specific configuration └── repository-guidelines.md # Project-specific constraints work/ # Multi-agent coordination ├── collaboration/ # Task orchestration and handoffs ├── reports/logs/ # Agent work logs └── notes/ # Exploratory work src/ # Production code tools/ # Development utilities tests/ # All test code 

📖 Read: Repository Quick Reference

1. Initialize your project

   # Fork this repository # Edit VISION.md and .doctrine-config/repository-guidelines.md

2. Review the doctrine stack

   • Read doctrine/DOCTRINE_STACK.md to understand the five-layer model
   • Browse doctrine/tactics/README.md to see available tactics
   • Review doctrine/guidelines/general_guidelines.md for core principles

3. Start working with agents

   • Use the specialist agents in doctrine/agents/
   • On GitHub, use the 'agents panel' (top-right icon) or assign issues to agents
   • Follow file-based orchestration via work/collaboration/

📖 See: Architectural Vision
📊 View: Doctrine Stack Diagram

Version 1.1.0 (2026-02-08):

• ✨ Doctrine Stack Integration: Five-layer governance framework (Guidelines → Approaches → Directives → Tactics → Templates)
• 🛠️ 9 Procedural Tactics: Stopping conditions, premortem, adversarial testing, AMMERSE, safe-to-fail, ATDD adversarial, test boundaries, input validation, code review
• 🔍 Hybrid Discovery: Directive-driven invocation + exploratory README discovery
• 📋 Updated Directives: 5 directives now explicitly invoke tactics (016, 017, 018, 021, 024)
• 📊 Architecture Diagrams: C4 PlantUML diagram for doctrine stack visualization
• 🎯 Dashboard Features: Real-time monitoring, file-based orchestration, WebSocket integration

📋 See: PR #132 - Doctrine Stack & Tactics Integration

This project draws inspiration from and acknowledges several pioneering works in AI-augmented development:

Source: lexler.github.io/augmented-coding-patterns
Author: @lexler

Comprehensive patterns and practices for integrating AI assistants into software development workflows. Essential reading for anyone building AI-augmented development systems.

Source: agents.md

Community-driven specification for AI agent behavior and interaction patterns. Provides standardized approaches for agent definition and orchestration.

Source: github.com/Priivacy-ai/spec-kitty
License: MIT
Contributions:

• Specification-Driven Development methodology (adapted) - Code serves specifications, not vice versa
• Dashboard interface design patterns - Live kanban board with WebSocket updates for real-time workflow visibility
• Multi-agent orchestration insights - Work package lane management and agent coordination patterns
• Template-based configuration approach - Quick-start templates with variable substitution and environment scanning
• Rich CLI feedback - Extensive use of the rich library for structured, colorful terminal output with progress indicators

We're grateful to the spec-kitty team for their innovative work in agent-augmented development. Their spec-driven.md guide and comparative architectural patterns have been particularly influential in shaping our workflow patterns and user experience enhancements.

Comparative Analysis: See docs/architecture/design/comparative_study/2026-02-05-spec-kitty-comparative-analysis.md for detailed analysis of adopted patterns.

Template Status: This repository serves as both a working example and a reusable template. Fork and adapt for your project needs.