AGETFramework GitHub

Knowledge that carries forward

Agents that
remember.
Work you can
govern.

AGET brings persistent knowledge, shared memory, and human supervision to repository-based AI agents. Keep context across sessions, coordinate a fleet, and build on the coding assistants you already use.

Your first session

Start with one
local trial.

See how AGET carries context before setting up a fleet. You’ll need Git, Python 3 with pip and venv, and a working coding assistant that can open a local repository.

  1. Make a local copy

    In a macOS or Linux terminal, run the commands below. Use a new folder and keep the default folder name for this trial.

    git clone https://github.com/aget-framework/template-supervisor-aget.git
    cd template-supervisor-aget
  2. Check the template

    From that folder, create an isolated Python environment, install the declared test dependency, and run the checks.

    python3 -m venv .venv
    . .venv/bin/activate
    python -m pip install -r requirements.txt
    python -m pytest tests/ -q

    Ready to continue: the test summary shows no failures. Some checks may be skipped. If a check fails or a tool is missing, pause and ask your coding assistant to explain the error before continuing.

  3. Open your first session

    Open this same folder in your coding assistant. Then send this message:

    Read AGENTS.md, then wake up.
    Explain what this template remembers between sessions.
    Do not change files.

    What to look for: the wake-up report names template-supervisor-aget and shows its version. This trial keeps the template’s identity; it is not yet your personalized supervisor.

Your assistant’s installation and account access are separate. Read the template setup guide

Explore AGET

Start small.
Build on what lasts.