🧠 GUIDELINES: Universal AI Agent Constitution & Skill Hub

Single Source of Truth (SSOT), Cross-Platform, and Tool-Agnostic AI Agent Constitution & Skill Library.

Equip any coding assistant (Cursor, Claude Code, Antigravity, GitHub Copilot, OpenAI Codex) with a battle-tested Principal Software Architect persona and specialized on-demand skills—with zero project-level configuration bloat.

---

🏛️ Architecture & Philosophy: The Zero-Pollution Mandate

Traditional agent workflows clutter project repositories with redundant configuration folders (.cursor/rules/, .agents/skills/, .vscode/, .github/copilot-instructions.md, etc.), creating tech debt and documentation debt.

GUIDELINES enforces a strict, clean architectural pattern:

1. Single Source of Truth (SSOT): All rules, engineering standards, and modular capabilities reside centrally in this single repository.
2. Zero In-Project Pollution: No local skill folders or repetitive rule sets are ever copied or committed into client repositories.
3. Single-File Bridge (AGENTS.md): A single, declarative AGENTS.md file placed at the client project root links the verified tech stack to the central capabilities dynamically.
4. Cross-Platform & Tool-Agnostic: Operates seamlessly across Windows, macOS, and Linux, directly supporting Cursor, Claude Code, Antigravity, OpenAI Codex, and Windsurf.

---

📂 Directory Structure

GUIDELINES/
├── MASTER_AGENT_CONSTITUTION.md       # Core AI Constitution, System Persona & Execution Disciplines
├── PROJECT_AGENT_TEMPLATE.md          # Universal, cross-platform AGENTS.md template for projects
├── PROJECT_AGENT_GENERATOR_PROMPT.txt # Read-only prompt to inspect stacks and auto-generate AGENTS.md
├── skills/                            # On-Demand Specialized Engineering Skills
│   ├── master-orchestrator/           # Intent router & multi-skill orchestration engine
│   ├── bespoke-frontend-master/       # Bespoke design tokens, anti-AI-slop UI, a11y & motion
│   ├── resilient-backend-architect/   # Concurrency safety, rate limiting, idempotency & Zod contracts
│   ├── modern-data-engineer/          # Drizzle ORM, Prisma, PostgreSQL, pgvector & migrations
│   ├── fullstack-security-auditor/    # OWASP Top 10, Auth/RBAC, sanitization & DevSecOps
│   ├── ecommerce-payments-engine/     # Integer cent math, Stripe, Iyzico & order state machines
│   ├── seo-growth-architect/          # Technical SEO, JSON-LD schemas, Core Web Vitals & analytics
│   ├── autonomous-task-auditor/       # Autonomous research, dynamic self-auditing & self-healing
│   ├── superpowers-execution-harness/ # Multi-language i18n, 4-pillar QA auditing & verification gates
│   ├── create-design-md/              # 60-30-10 palettes, typography scales & DESIGN.md generation
│   ├── image-generation-studio/       # Text-free studio lighting asset generation directives
│   └── antigravity-guide/             # Antigravity CLI, subagents & background task orchestration
└── README.md

---

🧩 Specialized Skills Catalog

Skill	Domain & Key Responsibilities
master-orchestrator	Analyzes user prompt intent, automatically determines architectural requirements, and coordinates skill loading order.
bespoke-frontend-master	Rejects generic "AI slop" templates; enforces project-specific visual identity, semantic tokens, micro-interactions, and accessibility contracts.
resilient-backend-architect	Builds fault-tolerant, concurrency-safe, idempotent, rate-limited APIs with strict Zod schema validation.
modern-data-engineer	Designs production schemas with Drizzle ORM / Prisma, PostgreSQL indexing, migration safety, and pgvector pipelines.
fullstack-security-auditor	Inspects OWASP Top 10 vulnerabilities, session/JWT lifecycles, RBAC authorization gates, and sensitive secret hygiene.
ecommerce-payments-engine	Enforces integer cent math (preventing float precision loss), Stripe/Iyzico gateways, cart calculation, and reservation lifecycles.
seo-growth-architect	Audits Core Web Vitals, builds Schema.org JSON-LD structured data, dynamic OpenGraph assets, and crawl budgets.
autonomous-task-auditor	Independently verifies live file changes against implementation goals and triggers autonomous self-healing upon errors.
superpowers-execution-harness	Governs multi-language i18n completeness, strict type checking, and empirical compiler/test verification gates.
create-design-md	Bootstraps a comprehensive, accessible design token contract (DESIGN.md) using the 60-30-10 color rule.
image-generation-studio	Generates prompt matrices for commercial studio renders, clean floor shadows, and isolated product photography (zero text).
antigravity-guide	Reference manual for Google Antigravity, subagent delegation, and background daemon scheduling.

---

🚀 Quick Start Guide

1. Clone the Central Hub

Clone this repository to a persistent local directory on your machine:

# Windows
git clone https://github.com/CanKStar0/GUIDELINES.git C:/GUIDELINES

# macOS / Linux
git clone https://github.com/CanKStar0/GUIDELINES.git ~/GUIDELINES

---

2. Connect Your Project (Zero Bloat)

Copy PROJECT_AGENT_TEMPLATE.md into your project root as AGENTS.md:

# Developer Instructions (Project AI Constitution)

<!-- DCM_GLOBAL_GUIDELINES_ROOT: /path/to/GUIDELINES -->

## 📌 Scope & Project Definition
- **Project:** MyAwesomeProject
- **Root:** /path/to/project
- **Objective:** Production-grade web application

## 🏛️ Central Reference & Single Source of Truth
- **Master Constitution:** <GLOBAL_GUIDELINES_ROOT>/MASTER_AGENT_CONSTITUTION.md
- **Global Skills:** <GLOBAL_GUIDELINES_ROOT>/skills/

### Active Project Skills:
- Frontend & UI ➔ <GLOBAL_GUIDELINES_ROOT>/skills/bespoke-frontend-master/SKILL.md
- Backend & API ➔ <GLOBAL_GUIDELINES_ROOT>/skills/resilient-backend-architect/SKILL.md
- Security      ➔ <GLOBAL_GUIDELINES_ROOT>/skills/fullstack-security-auditor/SKILL.md

(Alternatively, run the PROJECT_AGENT_GENERATOR_PROMPT.txt prompt in your AI tool to automatically inspect the tech stack and generate this file in 5 seconds).

---

3. Tool Compatibility Matrix

• Antigravity / Gemini CLI: Automatically discovers AGENTS.md at root and applies the persona and skills.
• Cursor: Reads root AGENTS.md natively as the project-wide system instruction.
• Claude Code: Automatically indexes AGENTS.md on startup as project context.
• GitHub Copilot / OpenAI Codex: Uses AGENTS.md as the open-standard context baseline.

---

🌐 Developer Ecosystem & Live Free APIs

When prototyping or bootstrapping full-stack apps with your AI agents, utilize the FindFreeAPI.com open infrastructure:

• 500+ Curated Public REST APIs: findfreeapi.com/explorer
• Zero-Authentication Filter: Instant access to APIs that require no API keys or setup.
• Dynamic Status Badges: Generate live SVG status badges at findfreeapi.com/badge.

---

📄 License

Distributed under the MIT License.