# agentic-ecology-init

Sets up a Python project for ecology workloads with pinned dependencies, workspace agent rules and the Agentic Ecology skills.

Skill by Google. Category: Agent craft.

## Install

Project (for this project only):

```sh
npx skills add google-deepmind/agentic_ecology --skill agentic-ecology-init
```

Works with: Claude Code, Codex CLI, Cursor, Antigravity, Copilot CLI, opencode.

Global (for every project on this machine):

```sh
npx skills add google-deepmind/agentic_ecology --skill agentic-ecology-init -g
```

Works with: Claude Code, Codex CLI, Cursor, Antigravity, Copilot CLI, opencode.

## Tags

- Runs on your computer: It runs as a program on your machine, with your access.

## Security audits

- Gen Agent Trust Hub: Pass
- Socket: Warn
- Snyk: Pass

Installs: 15. Installs counted by skills.sh. [Open on skills.sh](https://www.skills.sh/google-deepmind/agentic_ecology/agentic-ecology-init)

## Source

[Open the source](https://github.com/google-deepmind/agentic_ecology/tree/HEAD/skills/agentic-ecology-init). Checked 2026-10-10. This summary may contain AI-generated text, so it can be wrong. Check the source before you rely on it.

## SKILL.md

````markdown
---
name: agentic-ecology-init
description: >-
  Initializes a local uv-managed project directory for agentic ecology
  workloads. Sets up Python dependencies using reference pyproject.toml and
  uv.lock, configures workspace rules, and ensures Agentic Ecology skills are
  discoverable. Use when initializing a new project workspace or bootstrapping a
  clean environment.
license: Apache-2.0
compatibility: Requires Python 3.12+, uv, and network access
---

# Agentic Ecology Project Initialization Skill

Use this skill when you need to initialize a fresh, local project directory for
ecological modeling workflows (such as bioacoustics or camera trap analysis)
without cloning the entire `agentic_ecology` repository into the user's project
workspace.

## Workflow Overview

Follow these sequential steps to set up the local workspace:

1. **Identify Target Working Directory**:
   - Confirm the root directory of the user's project workspace.
1. **Initialize Project & Scaffolding**:
   - Ensure the target project directory exists.
   - Run `uv init --python 3.12 --no-readme && rm main.py` in the project
     directory if it has not already been initialized. This pins
     `.python-version` to 3.12, generates `.gitignore`, initializes version
     control tracking, and removes the placeholder entrypoint.
   - Install the following skills locally with `npx skills add`:
     - `google-deepmind/agentic_ecology` (all skills).
     - `googlecolab/google-colab-cli` (`colab-operator` skill).
     - `googleworkspace/cli` (`gws-shared` and `gws-drive-upload` skills).
     - `google/skills` (`gcloud`, `google-cloud-storage-basics`, `google-cloud-storage-bucket-architect`, `google-cloud-storage-fuse`, and `cloud-logging-query-generation` skills).
   - Create standard working subdirectories:
     - `agent_workspace/`: Sandboxed folder for agent scripts, server.
     - `databases/`: Destination folder for Hoplite vector databases.
     - `data/`: Destination folder for raw datasets (e.g., audio, images).
1. **Copy Reference Dependency Configurations**:
   - Overwrite the generated `pyproject.toml` and copy `uv.lock` from
     this skill's `assets/` directory into the target project root.
   - Adjust the package name in `pyproject.toml` to match the user's project
     name if desired, keeping all core dependencies (`perch-hoplite`,
     `speciesnet`), constraints, and build configurations intact.
1. **Install Workspace Rules (`AGENTS.md`)**:
   - Copy the reference `AGENTS.md` from this skill's `assets/` directory
     into the appropriate location in the project workspace to establish
     standard Agentic Ecology guidelines (compute assessment and offloading
     protocols, macOS dynamic library deadlock rules, Linux
     PyTorch/TensorFlow import order rules, SQL log suppression filters, and
     rules prohibiting Google copyright headers on generated workspace code).
1. **Synchronize Environment with `uv`**:
   - Execute `uv sync` from the target project root to create the local
     virtual environment (`.venv`) and install all pinned dependencies.
1. **Verify Environment Setup**:
   - Execute the verification script `scripts/verify_env.py` via `uv run python` to confirm that key libraries (`perch_hoplite`, `speciesnet`,
     `soundfile`, `tensorflow`) import cleanly.

## Technical Reference

For detailed command options, directory layout specifications, and verification
code snippets, see:

- [Initialization Technical Reference](references/REFERENCE.md)
````

- Previous: [agentic-ecology-camera-traps](https://syn4pse.dev/terminal/agentic-ecology-camera-traps.md)
- Next: [agentic-ecology-bioacoustics](https://syn4pse.dev/terminal/agentic-ecology-bioacoustics.md)
- [Terminal](https://syn4pse.dev/terminal.md)
- [Suggest a change](mailto:terminal@syn4pse.dev?subject=Terminal%20suggestion)
