oh-my-antigravity plugin
An Antigravity CLI plugin for engineering work: a separate design turn, documentation agents and edits only within a sealed plan
- 219 ★ on the repository
- Last commit
Broad access
Why this level
- The executor edits code and runs checks in your repository, though edits are limited to the sealed scope
- The plugin installs hooks in the Antigravity CLI and governs the agent's turns
Install
Manual install
agy plugin install https://github.com/Joonghyun-Lee-Frieren/oh-my-antigravityThen agy plugin list and /plugin list inside the CLI; Node.js 20 or newer is required.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The executor edits code and runs checks in your repository, though edits are limited to the sealed scope
- The plugin installs hooks in the Antigravity CLI and governs the agent's turns
This is third-party code. Review the repository files before installing.
What it does
oh-my-antigravity (OmA) is a plugin for the Antigravity CLI that runs an engineering task through a strict process. A design turn comes first: the Scope Brief records the objective, constraints, non-goals, decision owners and source conflicts, then the plan and checks are sealed and an independent auditor reviews that exact plan. Edits are allowed only in the next turn with /oma-team-exec, and /oma-team-verify runs real checks and separate section reviews. The plugin has 64 short skills, 53 of them /oma-<name> entrypoints, and 16 capability-limited agents (architect, planner, executor, verifier, documentation reader and others). Checkpoints are saved as immutable JSON and Markdown, resume notices drift between source and plan, and the evidence window gives bounded log excerpts. Missing services, permissions or evidence end in blocked or unknown status; a forced success is ruled out. Model recommendations live in config/models.json and were checked against Antigravity docs on 16 October 2026.
Who it is for. Antigravity CLI developers who want discipline: seal a plan, edit only within it and confirm results with real checks.
Good fit when
- A large change starts from a task description and acceptance criteria in /oma-team-plan
- Source documents must be read in full through /oma-read with a concise summary
- An independent auditor reviews the plan before the design turn closes
- Work runs through checkpoints and resumes after a break with a drift check
Not a fit when
- You need a one-line tweak: sealing a plan is too heavy
- You do not work in the Antigravity CLI: the plugin is built on its hooks and turns
- You want a zero-defect guarantee: the author says explicitly that there is none
Example request
/oma-team-plan Move authentication to access tokens, acceptance criterion: all authorization tests passLimitations
Requires a current Antigravity CLI and Node.js 20 or newer, plus Python 3.9 for Python files in the sealed scope. Models do not switch automatically: the native hooks cannot override the model, so profiles are only recommendations, and if the main model is unavailable pick the Flash profile by hand. Existing sealed plans must be resealed after upgrading to 1.2.0 for input protection. GitHub metadata shows no detected license, so read the terms in LICENSE.
How to disable. Run agy plugin uninstall for the plugin (list with agy plugin list). Checkpoints and reports live in the project working folder.
FAQ
What does the workflow look like?
First /oma-team-plan and a Scope Brief, then sealing the plan and an independent check, ending the design turn, then /oma-team-exec in the next turn and /oma-team-verify afterwards. On failures Stop schedules a repair batch.
What is Design Check?
A read-only agent's check of the exact sealed plan before the design turn can close. If the plan changes, earlier reviews are invalidated.
Which models does the plugin recommend?
The Balanced profile uses gemini-4-argon-high for design and review and gemini-3.8-flash-medium for implementation. The Flash profile swaps the design model for gemini-3.8-flash-high. Profiles are only recommendations; see agy models.
Does the plugin guarantee zero defects?
No, the author states this: safeguards do not guarantee zero defects or prove a model understood the delivered source. Where there is no evidence, the status stays unknown.
What are care profiles?
The routine, elevated and critical profiles set section sizes of 120, 90 and 60 lines and batches of 3, 2 and 1 units while keeping all checks and locks. Sensitive paths can only raise the level.
Discussion
No questions or reviews yet.
Related
PluginsA plugin for Antigravity CLI: the word lit at the start of a prompt starts a plan, row-by-row builds with tests, review, handoff
Installer of a working setup for Google Antigravity that puts rules, skills and workflows into a project with one npx command
Superpowers for Google Antigravity: a fork of the methodology with native agy tools and no translation layer.
Agentic OS adds a five-step workflow and checks in git hooks and CI so that a step without evidence does not count as done
A 33-skill plugin with gates: question the request, verify facts against sources, test first, review by a second model
Four Google Antigravity slash commands: bug investigation, a duel of two solutions, speed measurement and context compaction.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| oh-my-antigravity (this card) | no | 219 | |
| Antigravity CLI | yes | 2,790 | the Antigravity shell itself, while the plugin adds a workflow with plan sealing and checks |
| oh-my-claudecode | no | 40,789 | orchestration for Claude Code, while OmA is built on Antigravity CLI hooks and turns |
README in short
The README starts with installation and the first commands, then describes what is new in 1.2.0 (validator hashes, enforced source-read coverage, typed exits), the 1.1.0 features, model recommendations and a table of commands and roles. Details sit in docs/guide.