Run Google's Antigravity CLI (`agy`) from Claude Code to ask questions, delegate tasks, generate images, perform deep research, review git diffs, and select AI models per call.
Run a one-shot prompt through the Antigravity CLI and return its output verbatim
Delegate a task to the Antigravity (`agy`) runner subagent; supports background execution and model selection
Show all /agy:* commands, supported --model aliases, and canonical model names
Generate an image with Antigravity CLI (uses agy's built-in generate_image tool — Imagen under the hood)
Delegate a thorough research investigation to the agy:runner subagent
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Use Google's Antigravity CLI (agy) from
inside Claude Code. Delegate tasks to the agy:runner subagent, run quick
prompts, or get a second-opinion code review — without leaving your editor.
This plugin is for Claude Code users who already use (or want to start using)
Antigravity and want a smooth way to call it from the workflow they already
have. Intentionally small: no Node runtime, no broker, no review-gate hook —
just Bash and agy.
/agy:setup — verify agy is installed and authenticated; can install
it for you if it is missing./agy:ask [--model <alias>] <prompt> — one-shot prompt through agy -p;
returns the raw response./agy:delegate [--background] [--model <alias>] <task> — hand a task
to the agy:runner subagent. --background for long jobs./agy:research [--background] [--model <alias>] <topic> — delegate a
deep-research investigation; wraps the topic in a structured prompt and
routes through agy:runner./agy:image <description> — generate an image with agy's built-in
generate_image tool (Imagen under the hood). Optional --name and
--output./agy:review [focus] — ask Antigravity to review your current
git diff./agy:help — show all commands, supported --model aliases, and
canonical model names.agy:runner subagent — thin forwarding wrapper around the Antigravity
CLI; available as subagent_type: "agy:runner" for programmatic
delegation./plugin marketplace add …).agy) installed locally. /agy:setup can install it
on first run.agy: either OAuth cached in the system keyring (after one
interactive run of agy) or ANTIGRAVITY_API_KEY exported in your shell.PATH. macOS, Linux, or WSL.In Claude Code, run these three slash commands in order:
/plugin marketplace add simplybychris/antigravity-plugin-cc
/plugin install agy@antigravity-cc
/reload-plugins
Then verify everything is wired up:
/agy:setup
If agy is missing, /agy:setup offers to install it via the official
installer:
curl -fsSL https://antigravity.google/cli/install.sh | bash
If agy is installed but not logged in, run agy once interactively in your
terminal to complete OAuth — or export ANTIGRAVITY_API_KEY.
/agy:ask explain the difference between Go channels and Rust async in one paragraph
Returns Antigravity's response verbatim.
agy:runner subagent/agy:delegate refactor the SQL queries in src/db/queries.go to use prepared statements
For long tasks, run in the background and let Claude Code notify you when it finishes:
/agy:delegate --background investigate why integration tests are flaky in CI
You can also delegate by talking to Claude:
Ask agy to look at this file and suggest a simpler design.
The plugin's selection rules route through the agy:runner subagent
automatically.
Stage or make some changes, then:
/agy:review
/agy:review focus on error handling and concurrency safety
/agy:delegate --model sonnet fix the off-by-one in pagination
/agy:delegate --model pro write a high-coverage test for the cache layer
/agy:ask --model opus "explain Go's escape analysis"
Supported aliases: flash-low, flash-medium, flash, pro-low, pro,
sonnet, opus, gpt-oss. The canonical TUI strings (e.g.
"Claude Opus 4.6 (Thinking)") are also accepted. Run /agy:help for the
full table.
If no --model is given, the wrapper uses whatever the TUI is currently set
to — stored in ~/.gemini/antigravity-cli/settings.json. Project-local
AGENTS.md and GEMINI.md files are read directly by agy and unaffected
by this plugin.
/agy:research what's the current state of WebGPU support across browsers in 2026?
/agy:research --background --model opus survey post-quantum signature schemes used in TLS
The command wraps your topic in a research-oriented preamble (background,
key findings, caveats, sources) and delegates to agy:runner. Long
investigations work well in --background.
/agy:image a minimalist dark-mode login mockup, blue accent color
/agy:image --name hero --output ./assets/hero.png isometric illustration of a developer at a desk
Triggers agy's built-in generate_image tool. The image is written to
the Antigravity artifacts dir (e.g.
~/.gemini/antigravity-cli/brain/<uuid>/<name>.png). Pass --output if
you want the wrapper to copy it next to your project.
npx claudepluginhub simplybychris/antigravity-plugin-cc --plugin agyComprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Standalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.
Write feature specs, plan roadmaps, and synthesize user research faster. Keep stakeholders updated and stay ahead of the competitive landscape.
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.