autoctx-init

Installation
SKILL.md

Auto Context Generation Initialization

This skill guides the user through scaffolding a new environment for the DB Data Agent auto context generation. It handles creating tracking files, output directories, and configuring database connections.

Initialization Workflow

Follow these steps when the user asks to initialize the environment:

  1. Confirm Working Directory: Explicitly state the current working directory to the user. Explain that the initialization will create an autoctx/ folder in this directory to hold tools.yaml, state.md, and experiments/. Ask them to confirm if this is the correct location before proceeding.
  2. Check Existing Infrastructure:
    • Check if the autoctx/ directory exists.
    • If it exists, verify if it contains valid tools.yaml and state.md files. If it appears to be an unrelated folder or corrupted, STOP and ask the user how to proceed (e.g., use a different name or overwrite).
    • If it is a valid Autoctx folder and contains all items, inform the user it's already initialized. Otherwise, proceed to create missing items inside autoctx/.
  3. Setup Toolbox Configuration: If tools.yaml is missing inside autoctx/, follow the primary "1. Create a New tools.yaml" workflow documented below in the Toolbox Config Helper section.
  4. Create State Tracker: If state.md is missing inside autoctx/, create it with header “context authoring experiment state tracking”.
  5. Initialize Experiments Directory: If experiments/ is missing inside autoctx/, create an empty experiments/ directory inside autoctx/.

Output

Installs
1
GitHub Stars
33
First Seen
Jun 10, 2026
autoctx-init — googlecloudplatform/db-context-enrichment