unity-cli
An agent skill by Unity-Technologies, from Unity-Technologies/unity-agent-plugin. Tags: automation, cli, development, devops, editor-tools.
What it does
Use when interacting with Unity CLI from the terminal, or to control a running/connected Unity Editor from the command line — create or modify GameObjects, edit scenes and assets, inspect the hierarchy, and run C# in a live Editor instead of hand-editing scene or asset files. Also install, upgrade or uninstall editors, create, list or open projects, manage modules, manage licenses, check auth status, read logs, browse Unity releases, build/test projects, configure the Unity MCP server for AI agents, or run any other Unity CLI operation. For a guided idea-to-running-project flow for a brand-new game, use the new-unity-project skill instead.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add Unity-Technologies/unity-agent-plugin --skill unity-cli
Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/Unity-Technologies/unity-agent-plugin
cp -r unity-agent-plugin/skills/unity-cli ~/.claude/skills/unity-cli
Safety box score
Grade F, high risk: 81 of 100 risk points, rated 2026-10-19. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This skill controls a live Unity Editor, installs or configures Unity tooling, and reaches into auth/license/cloud/VCS/collaboration/Jira flows. The main risks are shell execution, credential handling, and broad destructive or externally visible changes; several referenced command guides were not read, so the full surface is not completely visible.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 3 (high) | It runs shell commands and can execute arbitrary C# against a live Editor (`unity command eval`, `unity shell --protocol ndjson`), and the install path includes `curl -fsSL https://unity.com/install.sh | bash`. |
| Network access | 3 (high) | It downloads installers and talks to Unity cloud, collaboration, Jira, licensing, and update/telemetry endpoints, so outbound data flow is core to the skill. |
| Filesystem reach | 3 (high) | It writes across project trees and user/system locations (`ProjectSettings/...`, `~/.local/bin`, keyrings, AI client skill dirs) and can delete or replace editor modules and worktrees. |
| Secrets & credentials | 3 (high) | It uses the auth broker/keyring and accepts secrets via env/stdin (`UNITY_SERVICE_ACCOUNT_SECRET`, `--secret-from-stdin`, `--git-token-stdin`, `--key-from-stdin`), so credential access is part of normal operation. |
| Privilege & persistence | 2 (moderate) | It installs software and changes global/system settings (`apt`/`dnf`-style installs, `/usr/share/keyrings`, `git config --global`, Windows UAC paths), but does not inherently require sudo for every action. |
| Instruction hijack surface | 1 (low) | The docs repeatedly instruct prompt suppression (`--yes`, `--non-interactive`, `--accept-eula`) in install/delete/return flows; that is bounded CI automation, not hidden instruction-following. |
| Irreversible actions | 3 (high) | It can create, sync, switch, resolve, publish, delete, or return resources across Unity projects, VCS, licenses, cloud orgs, collaboration threads, and Jira; many actions are externally visible or irreversible. |
| Transparency | 2 (moderate) | Several referenced command guides were not read (`build-run-test.md`, `diagnostics-maintenance.md`, `integration-advanced.md`, `projects-templates.md`, `CHANGELOG.md`), so part of the command surface remains unreviewed. |
Before installing
- Read `SKILL.md` around install/auth/license/VCS flows and decide if pipe-to-shell plus live C# execution is acceptable.
- Review `references/auth-license-cloud.md` for keyring, service-account, and stdin-secret handling.
- Inspect the unread guides (`integration-advanced.md`, `projects-templates.md`, `build-run-test.md`) before trusting the full command surface.
- Try `unity auth status --format json` and `unity --help` in a scratch account with secrets unset first.
Rated 2026-10-19 by an automated review of 9 files from the skill's repository.
Source
- Repository
- Unity-Technologies/unity-agent-plugin (all skills from this repository)
- Path
- skills/unity-cli/SKILL.md
- Branch
- main
- Updated
- 2026-09-22
Related skills
- bambu-labs — Dry-run, upload, and cautiously initiate local Bambu Lab print jobs from validated plain `.gcode`, using Bambu LAN FTPS/MQTT handoffs.
- devops-engineer — Creates Dockerfiles, configures CI/CD pipelines, writes Kubernetes manifests, and generates Terraform/Pulumi infrastructure templates.
- gcode — Generate, inspect, dry-run, and statically validate plain FDM `.gcode` from 3D mesh files by orchestrating real slicer CLIs. Use when Codex needs to slice.
- kubernetes-specialist — Use when deploying or managing Kubernetes workloads.
- vercel-cli-with-tokens — Deploy and manage projects on Vercel using token-based authentication.
- agents-deploy — Use when deploying your agent to AWS, or when a deploy has failed.