tools

A centralized, tool-agnostic developer intelligence and automation hub.

Overview

This repository consolidates AI behavior definitions, automation scripts, team processes, and stable reference knowledge into one place. It is intentionally model- and tool-agnostic so that its contents remain useful regardless of which LLMs, CI systems, or editors are in use.

Folder Structure

Folder	Purpose
ai/	Definitions that shape LLM behavior: agents, prompts, instructions, and skills
automation/	Executable scripts, CLIs, and CI pipeline configuration
process/	Human and system workflows for onboarding, releases, and code review
knowledge/	Stable reference material: architecture decisions, patterns, and conventions

Rules

• ai/ affects LLM behavior and must remain model/tool agnostic. Do not include provider-specific syntax or SDK calls.
• automation/ contains executable scripts and CLIs. Every entry point must be runnable without manual setup steps beyond what is documented.
• process/ contains human/system workflows. Keep steps concrete and actionable.
• knowledge/ contains stable reference material. Avoid content that changes frequently; prefer linking to authoritative sources instead.

Usage

Clone the repository

git clone https://github.com/utily/tools.git
cd tools

Run scripts in automation/

Browse automation/scripts/ or automation/cli/ for available scripts. Each subfolder contains a README.md describing what is available and how to invoke it.

# Example – run a script directly
bash automation/scripts/<script-name>.sh

Extend ai/ with new agents or prompts

1. Choose the appropriate subfolder (agents/, prompts/, instructions/, or skills/).
2. Add a new file following the naming convention used in existing files.
3. Document the new entry in the subfolder's README.md.
4. Open a pull request (see Contribution Model below).

Contribution Model

• All changes are submitted via pull request; direct commits to the default branch are not permitted.
• Keep the separation between top-level folders strict. An ai/ file must not contain runnable shell code; an automation/ file must not contain raw prompt text.
• Prefer small, focused pull requests that touch one folder at a time.
• Update the relevant subfolder README.md when adding or removing files.