google-gemma/gemma-skills

Skills for the Gemma and model/agent interactions

What it solves

It provides a set of specialized knowledge modules, called "skills," designed to improve how AI coding assistants and agents interact with the Gemma family of models. It bridges the gap between general-purpose AI assistants and the specific technical requirements for developing with or training Gemma.

How it works

These skills are distributed as plugins or modules that can be installed via various CLI tools (such as Vercel skills, Context7, or Antigravity). Once installed, they provide the AI agent with the specific context and capabilities needed to perform tasks related to Gemma.

Who it’s for

Developers and machine learning engineers who use AI coding assistants to build applications with Gemma or to train and fine-tune models on local hardware.

Highlights

  • gemma-dev: A skill focused on building applications and answering general knowledge inquiries about Gemma models.
  • gemma-trainer: A skill focused on the technical process of training, fine-tuning, and adapting models using techniques like SFT, DPO, RLHF, and Reward Modeling.

Related

  • Project
    QwenLM/Qwen-AgentWorld

    A native language world model that simulates agentic environments across seven domains to provide scalable and controllable training and evaluation for AI agents.

  • Project
    tensorlakeai/tensorlake

    Tensorlake is a cloud platform that offers fast, scalable MicroVM sandboxes for running AI agents safely. It provides a Python SDK and CLI to create, manage, snapshot, and pool sandboxes, a versioned object‑store filesystem, and a serverless orchestration layer where functions (decorated with `@function`) run in isolated sandboxes and can be invoked via HTTP.

  • Project
    melandlabs/openloomi

    OpenLoomi is an open‑source, cross‑platform desktop “attention agent” that continuously gathers context from your work tools (email, chat, issue trackers, docs) into a searchable, encrypted memory. It surfaces reminders and lets you ask an LLM for help directly inside messaging apps. The runtime is agent‑agnostic, offering Skills and Plugins that integrate with Claude Code, Codex, OpenCode, Hermes, OpenClaw, etc. Installable via simple binaries, it emphasizes local‑first privacy and works out‑of‑the‑box.

  • Project
    AgriciDaniel/compass

    A complete Obsidian vault template that transforms the app into a Life OS, integrating structured productivity workflows with an AI assistant and MCP bridge for automated life management.