linux-setup

Useful linux setup for developers.

Installation

This will update system packages

sudo apt update
sudo bash install-basic-package.sh

AI Agents Development

worktrunk via curl

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/max-sixty/worktrunk/releases/download/v0.24.1/worktrunk-installer.sh | sh && wt config shell install

Prompt

Starship: curl -sS https://starship.rs/install.sh | sh -s -- -b ~/.local/bin and add eval "$(starship init bash)" to .bashrc

Developer Utils (bash / Linux)

### Ubuntu/Debian package
sudo apt install fzf zoxide fd-find ripgrep bat shellcheck yazi

# additional utils needed for chafa
sudo apt install ffmpeg 7zip imagemagick

# install bun (or via npm: npm install -g bun)
curl -fsSL https://bun.com/install | bash

### fzf (download binary release) to ~/.local/bin
wget -qO- https://github.com/junegunn/fzf/releases/download/v0.67.0/fzf-0.67.0-linux_amd64.tar.gz | tar -xz -C ~/.local/bin

### zoxide (z / autojump inspired cd)
curl -sSfL https://raw.githubusercontent.com/ajeetdsouza/zoxide/main/install.sh | sh
# add this to .bashrc: `eval "$(zoxide init bash)"' >>~/.bashrc`


### fd (find alternative)
ARCH=$(uname -m) && wget -qO- "https://github.com/sharkdp/fd/releases/download/v10.3.0/fd-v10.3.0-${ARCH}-unknown-linux-gnu.tar.gz" | tar -xz -C ~/.local/bin --strip-components=1 --wildcards '*/fd'

### ripgrep
ARCH=$(uname -m) && wget -qO- https://github.com/BurntSushi/ripgrep/releases/download/15.1.0/ripgrep-15.1.0-${ARCH}-unknown-linux-musl.tar.gz | tar -xz -C ~/.local/bin --strip-components=1 --wildcards '*/rg'

### bat
ARCH=$(uname -m) && wget -qO- https://github.com/sharkdp/bat/releases/download/v0.26.1/bat-v0.26.1-${ARCH}-unknown-linux-gnu.tar.gz | tar -xz -C ~/.local/bin --strip-components=1 --wildcards '*/bat'

### shellcheck
ARCH=$(uname -m) && wget -qO- "https://github.com/koalaman/shellcheck/releases/download/stable/shellcheck-stable.linux.${ARCH}.tar.xz" | tar -xJv -C ~/.local/bin --strip-components=1

### ast-grep (via npm) - alternative: uv tool install (see below)
# npm install --global @ast-grep/cli

### ast-grep, ruff, ty, prek, zizmor
# curl -LsSf https://astral.sh/uv/install.sh | sh
# for tool in ast-grep-cli ruff hatchling prek zizmor; do uv tool install $tool; done

uv tool install ast-grep-cli
uv tool install ruff
uv tool install hatchling
uv tool install prek
uv tool install zizmor
uvx ty check

### actionlint
bash <(curl https://raw.githubusercontent.com/rhysd/actionlint/main/scripts/download-actionlint.bash) && mv actionlint ~/.local/bin/

### For Node based development:
npm install -g oxlint oxfmt vitest eslint-plugin-import
npm install -g typescript

AI Agent

Popular choices (both offline and online)

# kilo
npm install -g @kilocode/cli

# Claude Code
curl -fsSL https://claude.ai/install.sh | bash

# Gemini CLI
npm install -g @google/gemini-cli

# OpenAI Codex
npm i -g @openai/codex

Kilo

Install npm install -g @kilocode/cli Create ~/.config/kilo/opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ollama": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Ollama",
      "options": {
        "baseURL": "http://localhost:11434/v1"
      },
      "models": {
        "gpt-oss": {
          "name": "gpt-oss"
        }
      }
    }
  }
}

OpenCode + Ollama + gpt-oss

Install curl -fsSL https://opencode.ai/install | bash Create ~/.config/opencode/opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ollama": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Ollama",
      "options": {
        "baseURL": "http://localhost:11434/v1"
      },
      "models": {
        "gpt-oss": {
          "name": "gpt-oss"
        }
      }
    }
  }
}

Claude Code + Ollama + gpt-oss

Install curl -fsSL https://claude.ai/install.sh | bash Create claude-ollama.sh

#!/bin/bash
DEFAULT_OLLAMA_HOST="localhost"
DEFAULT_MODEL="gpt-oss"
DEFAULT_ANTHROPIC_BASE_URL=$DEFAULT_OLLAMA_HOST
if [ -z "$1" ]; then
    ANTHROPIC_AUTH_TOKEN=ollama ANTHROPIC_BASE_URL=$DEFAULT_ANTHROPIC_BASE_URL claude --model "$DEFAULT_MODEL"
else
    ANTHROPIC_AUTH_TOKEN=ollama ANTHROPIC_BASE_URL=$DEFAULT_ANTHROPIC_BASE_URL claude --model "$@"
fi

OpenAI Codex + Ollama + gps-oss

Install npm i -g @openai/codex and then run codex --oss -m gpt-oss:20b or create ~/.codex/config.toml

model = "gpt-oss:20b"
model_provider = "ollama"

[model_providers.ollama]
name = "Ollama"
base_url = "https://ollama.com/v1"
env_key = "<<OLLAMA_API_KEY>>"

Cline

After installing node 24, install npm install -g cline

using Qwen3.5 35B-A3b FP8 vllm w/ nvidia (<100 GB)

services:
  vllm-qwen:
    image: vllm/vllm-openai:nightly
    container_name: vllm-qwen3.5
    runtime: nvidia
    ports:
      - "11435:8000"
    volumes:
      - ~/.cache/huggingface:/root/.cache/huggingface
    environment:
      - HF_TOKEN=${HF_TOKEN}
    ipc: host
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    # The actual model launch command
    command: >
      --model Qwen/Qwen3.5-35B-A3B-FP8
      --kv-cache-dtype fp8_e4m3
      --max-model-len 131072
      --trust-remote-code
      --reasoning-parser qwen3
    restart: unless-stopped