hosamzewain/ai-assisted-development-framework

★ 18Audience · developerComplexity · 1/5Setup · easy

Mindmap

mindmap
  root((ai-dev-framework))
    What it does
      Shared AI context
      Team consistency
      Template repo
    Contents
      Context documents
      Coding standards
      Named skill files
      Glossary
    Use Cases
      Align AI behavior
      Auto-generate drafts
      Onboard new tools
    Rules
      No secrets in files
      Keep files minimal
      Human reviewed

mindmap root((ai-dev-framework)) What it does Shared AI context Team consistency Template repo Contents Context documents Coding standards Named skill files Glossary Use Cases Align AI behavior Auto-generate drafts Onboard new tools Rules No secrets in files Keep files minimal Human reviewed

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Set up your team's codebase with shared AI context files so every developer's Claude Code or Cursor session follows the same architecture and coding rules

USE CASE 2

Use the included prompt to have an AI scan your existing project and generate a draft set of context and skill files that your team then reviews

USE CASE 3

Add named skill files to guide your AI through specific repeated tasks like writing backend changes, fixing bugs, or reviewing pull requests

Tech stack

Markdown

Getting it running

Difficulty · easy Time to first run · 30min

In plain English

This repository is a template for software teams who want to use AI coding tools consistently across a shared codebase. The core problem it addresses is straightforward: AI assistants do not remember previous conversations, so if every developer is prompting their AI tool differently, the AI ends up making inconsistent decisions about architecture, testing, security, and style. This template is a proposed fix for that. The idea is to put all the project rules, architecture notes, and workflow instructions directly inside the repository itself, in files that any AI tool can read before it starts working. Instead of each developer explaining the project from scratch every session, the AI reads a small set of shared documents that describe what the product does, how the codebase is organized, and which rules must not be broken. This makes the AI's behavior more predictable across the team. The template ships with a folder structure that includes context and architecture documents, a set of coding and security standards, a testing policy, a release policy, a code review guide, and a glossary for product-specific terms. It also includes a set of named skills, each of which is a short instruction file guiding the AI through a specific type of task, such as making a backend change, fixing a bug, or reviewing a pull request. The approach is intentionally minimal: the AI reads a small always-loaded context file and only pulls in the other documents when the task calls for it. To use it, a team copies the template files into their own project repository and edits them to reflect their actual tech stack, folder layout, testing commands, and security model. There is also a prompt file that lets you ask an AI tool to inspect your codebase and generate a draft version of these files automatically, which the tech lead and senior developers then review and correct before it becomes official. The README is explicit about what not to include: no secrets, no API keys, no customer data, and no large requirement documents. The goal is a lean, maintainable set of files that keeps AI tools aligned without becoming a maintenance burden.

Copy-paste prompts

Prompt 1

I want to use the ai-assisted-development-framework template in my existing Node.js project. Which files do I copy first and what do I edit to reflect my actual tech stack and folder structure?

Prompt 2

Generate a draft CONTEXT.md for my project using this framework's format. Our stack is React, Node.js, and PostgreSQL. Include architecture overview, folder layout, and the three rules that must never be broken.

Prompt 3

I want to create a skill file for code reviews using this framework. What format should it follow and what belongs in the always-loaded context file vs an on-demand skill file?

Prompt 4

My team uses both Claude Code and Cursor. How does the ai-assisted-development-framework handle tools that have different skill and instruction file formats?

Open on GitHub → Explain another repo

← hosamzewain on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.