A curated list of tools, frameworks, and methodologies for the AI-enabled engineering team.
This repository tracks the rapidly evolving landscape of AI development tools. It moves beyond simple "chatbots" to focus on Agentic Workflows, Spec-Driven Development, and Developer Enablement.
- If you are a team lead:
- Use this repo to standardize how your team sets up AI-enabled services.
- Start with
AGENTS.mdfor agent behavior anddocs/standards/for language-specific guidelines. - Pick a collection under
collections/(for examplecollections/typescript-web-service.md) as the default stack for new projects.
- If you are an individual developer:
- Use the sections below to choose tools for your IDE, stack, and workflows.
- When creating a new service, follow the relevant standards in
docs/standards/.
- If you are an AI coding agent:
- Always read
AGENTS.mdfirst. - For new subprojects, follow the language standards in
docs/standards/and the relevant collection undercollections/.
- Always read
- Language and service standards:
docs/standards/typescript.md– TypeScript/Node service layout, tooling, tests, and error handling.docs/standards/python.md– Python service layout, tooling (pytest, ruff, mypy), and error handling.
- Ready-made stacks (collections):
collections/typescript-web-service.md– Opinionated stack for a TypeScript/Node web service with strict typing, Vitest, ESLint, and CI outline.
- Workflow guides:
docs/guides/choosing-mcp-servers.md– How to choose MCP servers based on your tech stack (Java/GitLab, Python/GitHub, .NET/Azure DevOps, etc.).
As you add more standards or collections, list them here so both humans and AI agents can discover them easily.
Tools that integrate directly into the coding workflow.
- GitHub Copilot - The industry standard for in-editor code completion and chat.
- Cursor - An AI-native fork of VS Code. It indexes your entire codebase for superior context awareness.
- GitLab Duo - The integrated AI suite for the GitLab ecosystem (DevSecOps).
- Windsurf - The "Agentic IDE" by Codeium. Uses "Flows" to maintain deep context across complex refactors.
- Kiro - An agentic AI development environment.
- Claude Code / Open Code - Command-line interfaces that allow LLMs to perform file operations and edits directly from the terminal.
- Gemini CLI - A command-line interface for interacting with the Gemini AI platform.
- Aider - The benchmark for AI pair programming in the terminal. Excellent at git-aware commits and refactoring.
- OpenAI Codex - A lightweight coding agent from OpenAI that runs locally in your terminal and integrates with ChatGPT and IDEs.
- Goose - A local-first AI engineering agent that automates complex tasks using MCP servers, extensions, and recipes.
- Kiro CLI - A command-line interface for interacting with the Kiro AI platform.
How we define software before the AI writes it. This is the shift to Spec-Driven Development (SDD).
- Agents.md - A proposal for a standard file format to document context/instructions for AI agents within a repository.
- GitHub Spec Kit - Tools and templates for defining structured specifications. This enables AI to generate high-quality code by strictly following a pre-defined "spec."
- OpenSpec - Spec-driven development (SDD) for AI coding assistants, with change proposals, tasks, and spec deltas that tools can drive via slash commands or AGENTS.md. See
docs/workflows/spec-driven-openspec.mdfor a full end-to-end example.
Frameworks for building your own internal agents and tools.
- Microsoft Agent Framework - A standardized way to build multi-agent conversations and workflows.
- AWS Bedrock Agents - Fully managed agents on AWS infrastructure.
- Microsoft Foundry - Comprehensive service for building and deploying AI apps on Azure.
- LangGraph - A library for building stateful, multi-actor applications with LLMs. The standard for complex orchestration.
- LangChain - A batteries-included framework for building LLM applications, with chains, tools, memory, and integrations across many providers.
- LlamaIndex - A data framework for connecting LLMs to your data, with indexing, retrieval, and RAG tooling.
- MCP (Model Context Protocol) - The new open standard for connecting AI assistants to systems (databases, Slack, GitHub) without building custom integrations for every model. See
docs/guides/choosing-mcp-servers.mdfor workflow-specific recommendations. - apm (Agent Package Manager) - A package manager for distributing and installing AI agent "skills" and tools.
- Agent Skills - An open format for packaging reusable agent skills as folders of instructions, scripts, and resources.
- n8n - Workflow automation tool that has pivotally embraced AI. Allows creating complex agentic workflows via a visual node interface.
Going from "Idea" to "Interface" in seconds.
- Lovable - Generates full-stack web apps from prompts. Excellent for MVPs.
- GitHub Spark - GitHub's platform for building and deploying intelligent apps, including AI-powered prototypes and workflows.
- v0 - Vercel's generative UI tool. Exports production-ready React/Tailwind code.
How do we trust what the AI built?
- Promptfoo - The CLI tool for testing LLM prompts. Essential for checking security jailbreaks, regressions, and output quality.
- LangSmith - Observability platform to trace, monitor, and debug agent workflows in production.
- Atlassian Rovo - An AI engine that connects data across the Atlassian suite (Jira, Confluence) to answer questions and manage requirements.
- NotebookLM - Google’s AI research and thinking partner that builds grounded assistants over your documents, transcripts, and notes.
- Ollama - Run Llama 3, DeepSeek, and other open-source models locally. Critical for testing agents without API costs or data leakage.
Found a new tool that changes the game? Open a PR to add it to the stack. If you add standards or collections, also:
- Update
docs/standards/orcollections/with clear, opinionated guidance. - Add links to new documents in the "Standards & Collections" section above so they are easy to discover.