Archives

Reproducing the devcontainer DX, but securely with SBX

Connect VS Code over SSH to an sbx sandbox built from kits, for a devcontainer-like workflow with microVM isolation, secrets protection and network policies.

Zed Editor + Docker Agent + ACP, but in a sandbox: running the agent with sbx

Run Docker Agent inside a Docker sandbox (sbx) and connect Zed to it over ACP, keeping llmman local while isolating the agent from your machine.

Zed Editor + Docker Agent + ACP: coding with a local agent plugged into llmman

Plug Zed's agent panel into Docker Agent over ACP, then point it at llmman to chat with a fully local, OpenAI-compatible coding agent.

llmman, the unified interface for language models

Install llmman, pull models like OCI images, and serve them with llama.cpp behind one OpenAI, Ollama and Anthropic compatible API.

Turbo Go, or 90s nostalgia

A Turbo Pascal-style IDE for Go in your terminal: install Turbo Go, write and run a first program, and tour its snippets, terminal, tree and LSP.

Docker Agent + llama.cpp: a local code agent in 5 minutes 🤖💙🦙

Serve a GGUF model with llama.cpp, point Docker Agent at it with a short agent.yaml, and get a fully local code agent running in minutes.

Bringing BeOS back to life: Haiku OS in a container, step by step

Run Haiku OS, the open source heir of BeOS, in a QEMU virtual machine inside a Docker container, and install it step by step from your browser.

Relearning Kotlin in lazy mode (with sbx)

Relearning Kotlin without installing anything: let an agent set up jbang, Kotlin and ten lessons inside an sbx sandbox, then turn that setup into a reusable kit.

Building a WASIP 2 component with TinyGo

Build a WASI Preview 2 component in Go with TinyGo: the WIT contract, the generated bindings, the wasip2 build, and calling it from Node.js through jco.

A mini code agent with Docker Agent - Part 9: a team of agents to spread the context out

Spread the context out with Docker Agent: a tool-less coordinator delegates to three Rust specialists, so every agent carries less and keeps room to think.

A mini code agent with Docker Agent - Part 8: a local agent to learn Rust

Assemble a local Rust learning companion with Docker Agent: a cookbook RAG, a rust-analyzer LSP toolset, and the context caps that keep a small model alive.

A mini code agent with Docker Agent - Part 7: `commands` (slash) + `instruction_file`

Make a local Docker Agent easier to drive: move the system prompt into a Markdown file, and add slash commands with argument interpolation.

A mini code agent with Docker Agent - Part 6: `skills`

Give a small local model detailed procedures on demand with Docker Agent skills: an on-disk SKILL.md, an inline one, and one forked into an isolated sub-agent.

A mini code agent with Docker Agent - Part 5: the `mcp` toolset

Plug external MCP tool servers into a local Docker Agent: a local stdio server, a remote streamable HTTP one, and a tools filter to keep a small model focused.

A mini code agent with Docker Agent - Part 4: the `rag` toolset

Ground a tiny local model's answers in your own documents: build a Rust cookbook RAG with Docker Agent, chunked embeddings and a local SQLite index.

A mini code agent with Docker Agent - Part 3: `think` + `memory` + `permissions`

Add reasoning, persistent memory and fine-grained command permissions to a Docker Agent mini code agent, with deny rules that survive yolo mode.

A mini code agent with Docker Agent - Part 2: `filesystem` + `todo`

Add typed filesystem operations and a todo task list to a local Docker Agent, with allow_list scoping to keep the small model safely inside its working directory.

A mini code agent with Docker Agent + Docker Model Runner - Part 1

Build a mini local code agent with Docker Agent and Docker Model Runner: a single shell tool on the small Mellum2 model, the agent loop, and running it safely in an sbx sandbox.

Docker Model Runner, standalone edition (`dmr`)

Install and use dmr, the standalone Docker Model Runner: one binary with the inference daemon and the full model CLI, no Docker Desktop needed.

Programmable guardrails for your local AI agent: NeMo Guardrails + Docker Agent, 100% on-device

Wire NVIDIA NeMo Guardrails into a local Docker Agent to catch jailbreaks, refuse out-of-scope requests, and mask PII on-device.