Antigravity Cli logo

Antigravity Cli

Community
scarletkc
antigravity-cli

Delegate bounded investigation, review, or implementation tasks to Antigravity CLI from a supervising agent. Includes a background client for task progress, follow-up prompts, conversation recovery, cancellation, and visible failures.

Overview

Publisherscarletkc
Repositoryagents
Skill nameantigravity-cli
Stars
225
Forks
12
Bundled files
3
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by scarletkc on GitHub. Read the source before you install it.

Installation

Install the Antigravity Cli AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/scarletkc/agents/tree/main/skills/antigravity-cli
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/scarletkc/agents.git /tmp/agents
mkdir -p .claude/skills
cp -r /tmp/agents/skills/antigravity-cli .claude/skills/antigravity-cli
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Antigravity Cli in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Antigravity Cli on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Antigravity Cli is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Antigravity CLI

Use the installed agy CLI as an external coding agent. The bundled client returns a task ID immediately and keeps a background worker available for follow-ups. It runs each turn with official streaming JSON input/output and resumes the exact Antigravity conversation on the next turn. Waiting for the process and stderr to finish makes permission denials and partial-output timeouts visible; each follow-up incurs CLI startup time.

Delegate when the user requests Antigravity or another agent can add value that you can verify. Keep routine work with the supervising agent when the extra startup and review time would outweigh that benefit.

Antigravity CLI must already be installed and authenticated. Run scripts/agy.py with the project's Python 3.12+ interpreter. The helper uses only the standard library; when the project has no Python environment, uv run --python 3.12 <script> ... also works. Resolve the script relative to this skill, not the project receiving the task.

Assign and follow up

Provide the goal, absolute working directory, relevant files and evidence, scope, and completion check. Antigravity does not inherit the supervising agent's conversation. Use a UTF-8 prompt file for substantial instructions:

sh
python <script> start --cwd <absolute-project-path> --prompt-file <task-file>
python <script> wait <task_id> --timeout 30
python <script> status <task_id>
python <script> reply <task_id> --prompt-file <follow-up-file>
python <script> cancel <task_id>
python <script> close <task_id>

Keep independent work with the supervising agent while the task runs. wait timing out only returns a snapshot; it does not cancel the task. reply is accepted after the current turn finishes. Retain the task_id and close the worker when no more follow-ups are needed.

The helper supplies both the process working directory and --add-dir. For concurrent edits to overlapping files, use separate worktrees and pass their paths as --cwd; the helper does not create worktrees. Check actual output paths, since Antigravity can also write to its conversation artifacts.

Configuration and permissions

Inherit Antigravity's authentication, model, and permission configuration. Subscription and API authentication remain Antigravity's responsibility. Do not switch providers, change authentication, or rewrite its global settings to conceal a failure.

start --model, --effort, --mode, and --sandbox are task-specific overrides. Honor the user's model or effort choice. Otherwise choose effort according to complexity, error cost, and ease of verification; omit the override when the model's support or the appropriate choice is uncertain. Discover models with agy --output-format json models.

Match permissions to the authorized work. --mode plan is a planning instruction, not enforced read-only access: headless plan review can proceed automatically. The supervising agent's sandbox does not configure AGY's. Do not promise isolation without checking the target platform's support.

For authorized workspace edits, --mode accept-edits allows file changes without interactive diff review; shell commands still follow tool permission rules. If a file tool rejects the target as an invalid artifact path, ask Antigravity to write a regular workspace file instead of an artifact. Inspect that tool error before attempting a shell workaround or expanding permissions.

Headless AGY cannot accept programmatic approval responses. When the helper reports blocked, inspect denied_actions and the partial result. If the task requires those tools, use scoped Antigravity permission rules or an interactive AGY session to establish the required permissions, following the original authorization. Do not repeatedly submit a blocked operation with unchanged permissions. The helper exposes --dangerously-skip-permissions only as an explicit override; do not use it merely to avoid an approval. Scoped rules limit which actions match; they are stored in Antigravity's global settings, not as temporary grants managed by this helper.

Check results and recover

Read status, closed, stop_reason, error, denied_actions, and tool_errors together. completed means the turn returned a response, not that its claims or edits were verified. Inspect the diff, output files, and relevant checks. AGY may recover from an individual tool error, so tool errors remain visible even when the final turn completes.

The helper does not equate native SUCCESS or exit code 0 with completion. It checks permission denials and stderr timeout diagnostics, and enforces its own startup, turn, and shutdown deadlines.

cancel stops that task's AGY process tree and closes its worker. It does not undo edits or external actions already performed. Read partial results before retrying. To continue a closed conversation, explicitly start a new task with its returned session_id:

sh
python <script> start --cwd <same-project-path> --conversation <session_id> --prompt-file <follow-up-file>

An open task can use reply after a provider failure. This is an explicit new turn in the same conversation; there is no automatic retry. If startup failed before a session ID was obtained, resolve the error and start again.

If installation, authentication, network, permissions, or execution blocks the task, tell the user what failed and what remains incomplete. Do not silently replace the requested Antigravity contribution with your own work or another provider. Committing, publishing, and other external actions retain the authorization boundaries of the original assignment.

Use <script> --help for commands and references/client.md for state fields, timeouts, storage, and diagnostics.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Antigravity Cli AI skill do?

Delegate bounded investigation, review, or implementation tasks to Antigravity CLI from a supervising agent. Includes a background client for task progress, follow-up prompts, conversation recovery, cancellation, and visible failures.

Why use Antigravity Cli on TypingMind?

Because you install it once and use it with any model. Antigravity Cli is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Antigravity Cli in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/scarletkc/agents/tree/main/skills/antigravity-cli. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Antigravity Cli?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Antigravity Cli?

As many as you like. As long as a model supports skills, you can use Antigravity Cli with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Antigravity Cli AI skill free?

Yes. It is published on GitHub by scarletkc under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇