MCP.Market

antigravity-cli skill

by NousResearch·NousResearch/hermes-agent·263k stars·MIT

Operate the Antigravity CLI (agy): plugins, auth, sandbox.

C70/100content scan

Is the antigravity-cli skill safe?

Read the findings before you install it. We read 2 files in the folder on 2026-10-25.

  • highreferences/cli-docs.md:9

    Downloads a script and runs it in one step, so what runs is whatever that server sends that day. Common for installers, and still worth a look at the address.

    - macOS/Linux: `curl -fsSL https://antigravity.google/cli/install.sh | bash`

Install the antigravity-cli skill

A skill is a folder. Copy it into your agent's skills folder and the agent loads it when the task matches its description.

git clone --depth 1 https://github.com/NousResearch/hermes-agent.git /tmp/hermes-agent
mkdir -p ~/.claude/skills
cp -r /tmp/hermes-agent/optional-skills/autonomous-ai-agents/antigravity-cli ~/.claude/skills/antigravity-cli
available in every project

In the Claude apps, zip the folder and upload it from the Skills settings. The folder on GitHub

The instructions your agent would load

SKILL.md as published, without the frontmatter. Read it on GitHub

Antigravity CLI (agy)

Operator guide for the Antigravity CLI, invoked as agy. Run all agy commands through the Hermes terminal tool; inspect its config and logs with read_file. This skill is reference + procedure — it does not wrap a network API, so there is nothing to authenticate from Hermes itself.

When to Use

  • Installing, updating, or smoke-testing the agy binary
  • Driving non-interactive agy --print / agy -p one-shots
  • Debugging Antigravity auth, sandbox, permissions, or plugin state
  • Reading Antigravity settings, keybindings, conversations, or logs

Mental model

Antigravity has two layers — keep them distinct or the guidance will be wrong:

agy update, agy changelog. Run these through the terminal tool.

  1. Shell wrapper commands — agy help, agy install, agy plugin,

/skills, /agents, etc. These only exist inside a running agy TUI session, not on the shell wrapper.

  1. Interactive in-session slash commands — /config, /permissions,

agy help shows the shell wrapper surface, NOT the in-session slash commands.

Prerequisites

command -v agy && agy --version.

  • The agy binary on PATH. Verify through the terminal tool:

auth via the OS keyring / browser sign-in (see Authentication below).

  • No env vars or API keys required by this skill — Antigravity manages its own

How to Run

Invoke every agy command through the terminal tool. Examples:

terminal(command="agy --version")
terminal(command="agy help")
terminal(command="agy plugin list")
terminal(command="agy --print 'Summarize the repo in 3 bullets'", workdir="/path/to/project")

For an interactive multi-turn TUI session, launch agy with pty=true (and tmux for capture/monitoring), the same pattern the codex / claude-code skills use. For one-shot smoke tests and scripted prompts, prefer agy --print (non-interactive).

To inspect Antigravity's own files, use read_file on the paths under Core paths below — do not cat them through the terminal.

Delegation patterns

agy is a coding-agent backend in the same family as codex / claude-code, so the same delegation shapes apply. Use these when handing real work (features, fixes, reviews, second opinions) to Antigravity rather than just smoke-testing.

One-shot (preferred for scripted prompts and second opinions)

terminal(command="agy -p 'Review this diff for bugs and security issues' --model 'Gemini 3.1 Pro (High)'", workdir="/path/to/repo", timeout=300)

-p is non-interactive: it runs the prompt and exits. Pick the engine with --model (run agy models for the exact display strings, e.g. 'Gemini 3.1 Pro (High)', 'Claude Opus 4.6 (Thinking)'). Add extra context roots with repeatable --add-dir.

Long / bounded runs (tests, builds, multi-file changes)

Background it and get notified on completion, the same as the codex skill:

terminal(command="agy -p 'Implement the change described in TASK.md and run the tests' --dangerously-skip-permissions", workdir="/path/to/repo", background=true, notify_on_complete=true)
# then: process(action="poll"/"log"/"wait", session_id=<id>)

Interactive multi-turn (PTY + tmux)

For a conversational session, launch agy -i (or bare agy) under pty=true with tmux for capture-pane / send-keys, exactly the pattern documented in the codex / claude-code skills. Resume later with --continue / -c or a specific --conversation .

Parallel instances (batch sub-issue / worktree fan-out)

Create one git worktree per task and launch an independent agy -p in each (background), then collect results — same worktree fan-out the codex skill uses for batch issue fixing. Bound concurrency to what the machine and your review capacity can absorb.

Output + bounding caveat (differs from Claude Code)

no result envelope with session_id / cost / turn count. Parse stdout directly; don't expect a JSON object.

  • agy -p returns plain text — there is no --output-format json and

(default 5m). Raise it for long tasks: --print-timeout 20m. Pair with the terminal timeout= so the outer call doesn't cut the run short.

  • There is no --max-turns. A print run is bounded by --print-timeout

Orchestration boundary

Antigravity is a worker execution backend or third-opinion reviewer — an execution detail owned by the agent/profile running a task, NOT a first-class orchestration primitive. Do not put agy on a kanban board as its own card or treat it as a coordination layer; route work through the normal task graph and let the assigned worker choose agy (vs. codex/claude-code/direct tools) as its method. Reach for it explicitly only when the user asks, when a worker is configured to wrap it, or when you want a Gemini-family cross-check against another agent's plan or diff.

Core paths

  • Binary / entrypoint: agy
  • App data dir: ~/.gemini/antigravity-cli/
  • Settings file: ~/.gemini/antigravity-cli/settings.json
  • Keybindings file: ~/.gemini/antigravity-cli/keybindings.json
  • Logs: ~/.gemini/antigravity-cli/log/cli-*.log
  • Conversations: ~/.gemini/antigravity-cli/conversations/
  • Brain artifacts: ~/.gemini/antigravity-cli/brain/
  • History: ~/.gemini/antigravity-cli/history.jsonl
  • Plugin staging: ~/.gemini/antigravity-cli/plugins//

Quick Reference

Wrapper commands

  • agy changelog
  • agy help
  • agy install
  • agy plugin / agy plugins
  • agy update

Useful flags

  • --add-dir
  • --continue / -c
  • --conversation
  • --dangerously-skip-permissions
  • --print / -p
  • --print-timeout
  • --prompt
  • --prompt-interactive / -i
  • --sandbox
  • --log-file
  • --version

Plugin subcommands (agy plugin --help)

enable , disable , validate [path], link , help

  • list, import [source], install , uninstall ,

Install flags (agy install --help)

  • --dir, --skip-aliases, --skip-path

In-session slash commands

/rename , /clear, /fork, /reset, /new

  • Conversation control: /resume (/switch), /rewind (/undo),

/keybindings, /statusline, /tasks, /skills, /mcp, /open , /usage, /logout, /agents

  • Settings & tools: /config, /settings, /permissions, /model,

not streaming), ! runs a terminal command directly, ? opens help

  • Prompt helpers: @ path autocomplete, esc esc clears the prompt (when

Settings and permissions

More skills from NousResearch/hermes-agent

  • A1passwordSet up op CLI, sign in, and read or inject secrets.
  • A3-statement-modelBuild integrated IS/BS/CF financial workbooks in Excel.
  • AaccelerateRun PyTorch training across GPUs with minimal changes.
  • Cactual-setupSet up Actual Computer (actual.inc) inference in Hermes.
  • Aadversarial-ux-testRoleplay a hostile user to find and triage UX pain points.
  • Aagent-merge-conflict-arbiterNeutral arbiter for merge conflicts between two agents.
  • AagentmailUse when an agent needs AgentMail CLI email inboxes.
  • Aai-presenter-videoMake a verified AI presenter video from script + image.
  • AairtableAirtable REST API via curl. Records CRUD, filters, upserts.
  • AanatomyExplain an idea as an interactive isometric machine figure.
  • Aapple-notesManage Apple Notes via memo CLI: create, search, edit.
  • Aapple-remindersApple Reminders via remindctl: add, list, complete.

All agent skills → · MCP servers