labs-ai-tools-for-devs

# AI Tools for Developers

Agentic AI workflows enabled by Docker containers.

Just Docker. Just Markdown. BYOLLM.

## MCP
Prompts and tools can be used as [MCP servers](https://www.anthropic.com/news/model-context-protocol) using the `--mcp` flag and registering prompts via git ref or path with `--register <ref>`.

![overall architecture diagram preview](img1.png)

Source for experiments in our [LinkedIn newsletter](https://www.linkedin.com/newsletters/docker-labs-genai-7204877599427194882/)

[**VSCode Extension**](https://github.com/docker/labs-ai-tools-vscode)

[**Docs**](https://vonwig.github.io/prompts.docs/)

# What is this?

This project provides a Docker image that enables workflows by combining Dockerized Tools, Markdown, and any LLM.

## Markdown is the language

Markdown is used to write complex workflows, which can be run with your own LLM in any environment, thanks to Docker.

## Dockerized Tools
![dockerized tools](img4.png)

Docker images are used as tools, enabling the LLM to take complex actions, get more context with fewer tokens, work across environments, and operate in a sandboxed environment.

## Conversation *Loop*
The conversation loop is the core of each workflow. Tool results, agent responses, and markdown prompts are passed through the loop. The agent can retry tools with different parameters until the desired result is achieved.

## Multi-Model Agents
Prompts can be configured to run with different LLM models, allowing you to use the best tool for the job and create multi-agent workflows.

## Project-First Design
The project provides the necessary context for the assistant in the software development loop.

### Extracting project context
![extractor architecture](img2.png)

Extractors are Docker images that extract information from a project into a JSON context.

## Prompts as a trackable artifact
![prompts as a trackable artifact](img3.png)

Prompts are stored in a git repo and can be versioned, tracked, and shared.

# Get Started
The VSCode extension is recommended for creating and running prompts with your own LLM. Alternatively, the CLI can be used with Docker.