k33g_org's Blog

I have always loved using devcontainer, because it lets VS Code open a project in a preconfigured container with its tools, dependencies and extensions, so you can start coding right away in a reproducible environment without installing everything on your…

In two minutes In the previous article, I showed you how to plug Zed into Docker Agent thanks to the ACP protocol, with llmman serving the model locally.

In two minutes Today I'm showing you how to plug Zed, my code editor, into an agent that runs locally, thanks to the ACP protocol.

llmman in two minutes llmman is a command-line tool written in Rust with the slogan "Run any agent on any model". llmman is a project initiated by Eric Curtin 👏 Its 3 main functions are as follows: It manages models like OCI images.

My first real IDE was Borland's Turbo Pascal editor (or maybe QBasic 🤓), and I have to admit I missed it a lot, especially when I'm in a terminal.

My favourite code agent is still Docker Agent, especially when I work with "local" LLMs, be they "big" (gemma-4-26B-A4B-it) when I'm on my work laptop, or more modest (Mellum2-12B-A2.5B-Instruct) when I'm on my personal Mac Book Air.

A few years ago (I think it was 1997 or 1998 😮) I had bought the installation floppies of BeOS, the x86 PC version (before BeOS R5).

The more I use AI, the more I'm torn about the way we developers use it.

WASI Preview 2 changes the foundations of WASI Preview 1 (where, for instance, a function can only exchange numbers).

And here we are at the 9th and final step of our Docker Agent series.

Over the previous seven steps, we saw various ways of using Docker Agent and discovered some of its basic features.

A useful agent is a reusable and drivable agent.

A skill is a procedure written in Markdown, of which only the name and the description are injected into the system prompt.

So far, all our tools were built into docker-agent (shell, filesystem, todo, think, memory, rag).

We have seen how to make an agent that can explore (01), work with files (02), think and stay on track (03).

Our agent can now explore (01) and work with files (02).

In the previous blog post, the agent only had shell.

Foreword At Docker, there are 2 open source projects I use every day: Docker Model runner (DMR): a local LLM model server behind an OpenAI-compatible API (but not only).

My favourite tool for serving local models is Docker Model Runner.

How to wire NVIDIA NeMo Guardrails into a docker agent agent to catch jailbreaks / prompt-injections, politely refuse out-of-scope requests, and mask the PII in tool results — all of it without a single byte of data leaving the machine, with a rules engine

The problem: as soon as an AI agent reads a file, PII (personally identifiable information: emails, IBANs, cards, names...) may travel to the model, that is, to its inference endpoint.

Content moderation with Llama Guard 2 This blog post is a follow-up to the previous article on ShieldGemma 2B, which showed how to plug a local safety classifier into a docker agent to moderate user prompts without ever sending data to a cloud service.

Content moderation with ShieldGemma How to add a 100% on-device content safety guardrail to a docker agent, by plugging Google's ShieldGemma 2B classifier into a simple hook and without any data ever leaving the machine.

Final tweaks before I start learning Rust.

Securing and adding knowledge One thing you can't skip when working with agents is running them in a secure environment.

For over 2 years now, I've been passionate about language models (LLMs) that you can run locally on your own machine, without depending on an external API.