Skip to content

slides.md

Latest commit

 

History

History
1668 lines (1155 loc) · 32.8 KB

slides.md

File metadata and controls

1668 lines (1155 loc) · 32.8 KB
Raw
theme seriph
background https://images.unsplash.com/photo-1451187580459-43490279c0fa?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80
addons
slidev-addon-qrcode
class text-center
highlighter shiki
lineNumbers true
info ## Agentic Coding with Antigravity CLI By Kenneth Kousen Learn more at [KouseniT](https://kousenit.com)
drawings
persist
transition slide-left
title Agentic Coding with Antigravity CLI
mdc true
slidev
slide-number controls progress
true
true
true
css unocss
<style> .slidev-page-num { display: block !important; opacity: 1 !important; visibility: visible !important; position: fixed !important; bottom: 1rem !important; right: 1rem !important; z-index: 100 !important; color: #666 !important; font-size: 0.875rem !important; } </style>

Agentic Coding with Antigravity CLI

Google's AI Agent in Your Terminal

Press Space for next page

Contact Info

Ken Kousen Kousen IT, Inc.

Course Overview

  • Duration: 5 hours of hands-on learning
  • Format: Instructor-led with multiple labs
  • Hands-on Labs: Real codebases in Python, JavaScript, Java
  • Prerequisites: Command-line experience, development background

Topics Covered

  • Foundation: Installation, CLI basics, authentication
  • Core Skills: File operations, shell integration, context management
  • Customization: AGENTS.md, custom commands, settings.json
  • Safety: Sandbox mode, permission rules, checkpointing
  • Advanced: MCP integration, skills/plugins, session management

What's New in Antigravity CLI 1.0.x

  • Initial 1.0.0 release of the Antigravity CLI — a Go-based terminal agent
  • Multi-model sessions: Gemini, Claude, and GPT-OSS via /model (1.0.5)
  • models subcommand + --model flag to pick a model at launch (1.0.5)
  • /permissions command to manage tool permission rules in-CLI (1.0.5)
  • G1 credits with /credits, /usage, /quota when quota runs out (1.0.3)
  • MCP via url in mcp_config.json; parallel server init (1.0.5/1.0.6)
  • SQLite conversation format + /resume, --continue, --conversation (1.0.4)
  • Plugin discovery for skills and subagents (1.0.1)
  • Latest stable track: Antigravity CLI 1.0.6

What is Antigravity CLI?

  • Google's Go-based AI coding agent for the terminal (binary: agy)
  • Multi-model: Gemini 3.x, Claude Sonnet/Opus 4.6, GPT-OSS — switch with /model
  • Built-in tools: web search, file ops, shell, web fetch
  • Model Context Protocol (MCP) support via mcp_config.json
  • Shares config with the Antigravity 2.0 desktop app (/export)

Key Differentiators

  • Model choice: mix Gemini, Claude, and GPT-OSS in one tool
  • Google Search Grounding: real-time web access
  • Free tier + G1 credits: keep working when quota runs out
  • Desktop continuity: shares conversations with Antigravity 2.0
  • MCP Native: built-in Model Context Protocol support

Models Available

  • Gemini 3.1 Pro: high-capability Gemini model for complex coding
  • Gemini 3.5 Flash: faster, lower-cost option (Low/Medium/High)
  • Claude Sonnet 4.6 / Opus 4.6: Anthropic models (thinking)
  • GPT-OSS 120B: open-weights option
  • List models: agy models · Switch: /model or --model

Quota & Credits

  • Free tier: sign in with a Google account to get started
  • G1 credits: kick in automatically when standard quota runs out
  • In-CLI panels: /credits, /usage, /quota for real-time status
  • Manage models and preferences via /settings

📖 Docs: antigravity.google/docs

Installation

  • macOS / Linux: curl -fsSL https://antigravity.google/cli/install.sh | bash
  • Windows (PowerShell): irm https://antigravity.google/cli/install.ps1 | iex
  • Installs the agy binary to ~/.local/bin/
  • Verify: agy --version · Update later: agy update
# Install (macOS / Linux)
curl -fsSL https://antigravity.google/cli/install.sh | bash

# Verify installation
agy --version

Authentication

  • Google Sign-In (default): launches automatically on first run
  • Remote / SSH: shows an authorization URL with a one-time code
  • API key (alternative): export ANTIGRAVITY_API_KEY="your-key"
  • Credentials persist via OAuth in ~/.gemini/
# First run: sign in with your Google account
agy

# Alternative: API key for scripts / headless use
export ANTIGRAVITY_API_KEY="your-api-key"
agy -p "Summarize this repo"

Basic Usage Modes

  • Interactive REPL: agy - Start a conversation
  • One-shot / print: agy -p "prompt" - Single response
  • Piped input: echo "task" | agy -p
  • Interactive with initial prompt: agy -i "initial context"
# Interactive mode
agy

# One-shot (print) mode
agy -p "Explain what this codebase does"

# Start interactively with an initial prompt
agy -i "You are a Python expert"

layout: image-right image: https://images.unsplash.com/photo-1555949963-ff9fe0c870eb?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Core Features

Essential Capabilities

Master the fundamentals

File References with @

  • Reference files directly: @./src/main.js
  • Reference directories: @./src/ (recursive)
  • Reference images: @./screenshot.png
  • Multiple references in one prompt
# Reference a specific file
agy -p "Explain @./src/app.py"

# Reference multiple files
agy -p "Compare @./old.js and @./new.js"

# Reference a directory
agy -p "Analyze the architecture in @./src/"

Shell Integration with !

  • Execute shell commands: !git status
  • Toggle persistent shell mode: !
  • The agent can observe and analyze output
  • Combine with AI analysis
# In interactive mode:
> !npm test
# The agent sees the test output

> !git diff
# Ask the agent to analyze the changes

# Toggle persistent shell mode
> !

Slash Commands: Navigation

  • /help - Show available commands and shortcuts
  • /context - View loaded context and token usage
  • /model - Switch the active model mid-session
  • /settings - Open settings and preferences

Slash Commands: Workflow

  • /agent <task> - Dispatch an asynchronous subagent
  • /permissions - Add/edit/remove tool permission rules
  • /usage - Session, quota, and rate-limit status

Slash Commands: Session Control

  • /resume - Open the conversation browser
  • /export - Send the session to the Antigravity 2.0 desktop app
  • /credits · /quota - Credit balance and quota panels

Slash Commands in Action (Context)

# Show what context is loaded and token usage
/context

# Switch the active model
/model

# Manage tool permission rules
/permissions

# Open settings and preferences
/settings

Slash Commands in Action (Sessions)

# Open the conversation browser
/resume

# Dispatch an async subagent task
/agent refactor the auth module

# Push this session to the desktop app
/export

Keyboard Shortcuts: Editing

  • Ctrl+L - Clear screen
  • Ctrl+V - Paste text/images
  • Ctrl+R - Open the Artifact Review panel

Keyboard Shortcuts: Control

  • Esc - Interrupt the active agent stream
  • Ctrl+C - Cancel current operation
  • Ctrl+D Ctrl+D - Exit Antigravity CLI (press twice)

Built-in Tools

  • File System: read_file(), write_file(), replace(), glob()
  • Shell: Execute terminal commands
  • Web: google_web_search(), web_fetch()
  • Memory: save_memory() for cross-session recall
# The agent automatically uses appropriate tools
"Search the web for React 19 new features"
# Uses google_web_search()

"Read all Python files in src/"
# Uses glob() and read_file()

"Update the README with the changes we made"
# Uses write_file()

layout: image-right image: https://images.unsplash.com/photo-1488590528505-98d2b5aba04b?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Safety & Control

Work Safely

Permission rules and sandboxing

Tool Permissions

  • Default: prompt for approval on each tool call
  • /permissions: add, edit, or remove allow/deny rules in-CLI
  • proceed-in-sandbox: auto-approve commands that stay in the sandbox
  • --dangerously-skip-permissions: auto-approve everything (use with care)
# Default - interactive, prompts per tool call
agy

# Auto-approve all tool calls (use with care)
agy --dangerously-skip-permissions

# Manage allow/deny rules from inside a session
> /permissions

Artifact Review

  • Review proposed changes before they are applied
  • Open the Artifact Review panel with Ctrl+R
  • Inspect diffs with /diff (supports commit-hash selection)
  • Works even while answering pending tool-permission prompts
  • Great safety habit when exploring an unfamiliar codebase
# Inspect changes from inside a session
> /diff

# Toggle the Artifact Review panel
Ctrl+R

Sandbox Mode

  • Isolate file operations away from your host
  • Multiple backends depending on your OS
  • Prevents accidental system changes
  • Perfect for exploring unfamiliar code
# Run in sandbox mode (terminal restrictions enabled)
agy --sandbox

# Combine with a one-shot prompt
agy --sandbox -p "Refactor this entire codebase"

Sandbox & Permission Modes

  • --sandbox: run with terminal restrictions enabled
  • proceed-in-sandbox permission mode: auto-approve commands that stay inside the sandbox, prompt only when one tries to break out
  • Sandbox isolation is enforced in headless print mode too (-p)
  • Pair with /permissions rules for durable guardrails
# Interactive, sandboxed
agy --sandbox

# Non-interactive, sandbox still enforced
agy --sandbox -p "Audit @./src for risky calls"

📖 antigravity.google/docs

Checkpointing

  • Snapshots: capture state before file modifications
  • Shadow storage: kept under ~/.gemini/ (not your repo)
  • Includes: files + conversation + tool call
  • Manage checkpoints from the /context panel
# View context and manage checkpoints
> /context

Restoring Checkpoints

# Open the context panel to review checkpoints
/context

# Resume an earlier conversation by ID
agy --conversation <id>

# Continue the most recent conversation
agy -c

Recover earlier state and pick up where you left off

layout: image-left image: https://images.unsplash.com/photo-1454165804606-c3d57bc86b40?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Context Management

AGENTS.md Files

Project memory and instructions

What is AGENTS.md?

  • Project memory loaded automatically
  • Coding standards and conventions
  • Architecture context for the AI
  • Persistent instructions across sessions
  • Antigravity also recognizes GEMINI.md and CLAUDE.md

Hierarchical Loading

graph TD
    A[~/.gemini/AGENTS.md<br/>Global Rules] --> B[project/AGENTS.md<br/>Project Standards]
    B --> C[project/frontend/AGENTS.md<br/>Frontend Conventions]
    B --> D[project/backend/AGENTS.md<br/>Backend Patterns]

    style A fill:#FFE5CC,stroke:#333,stroke-width:2px,color:#000
    style B fill:#CCE5FF,stroke:#333,stroke-width:2px,color:#000
    style C fill:#E5CCFF,stroke:#333,stroke-width:2px,color:#000
    style D fill:#CCFFE5,stroke:#333,stroke-width:2px,color:#000
Loading

More specific files override general ones

Example AGENTS.md

# Project: Weather API
## Tech Stack
- Backend: Python Flask
- Database: PostgreSQL
- Testing: pytest
## Coding Standards
- Use type hints for all functions
- Follow PEP 8 style guide
- Write docstrings for public APIs
## Current Focus
Implementing caching layer for API responses

Managing Context

  • Edit AGENTS.md directly to update project memory
  • /context - view combined context and token usage
  • Add includeDirectories in settings to load shared context
  • Restart or reload the session to pick up changes
# Create or edit project memory
$EDITOR AGENTS.md

# See what context is loaded
> /context

Modular Imports

  • Import other files with @file.md syntax
  • Break large context into components
  • Supports relative and absolute paths
# AGENTS.md

## Project Overview
@./docs/architecture.md

## Coding Standards
@./docs/style-guide.md

## API Documentation
@./docs/api-reference.md

layout: image-right image: https://images.unsplash.com/photo-1558494949-ef010cbdcc31?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Configuration

Customize Your Setup

settings.json and environment

Configuration Layers

  1. Default values - Built-in defaults
  2. User settings - ~/.gemini/settings.json
  3. Project settings - .gemini/settings.json
  4. Environment variables - Including .env files
  5. Command-line arguments - Highest priority

settings.json Options

{
  "general": {
    "preferredEditor": "vscode",
    "vimMode": false,
    "checkpointing": { "enabled": true }
  }
}

settings.json Options (UI + Tools)

{
  "ui": {
    "hideTips": false,
    "hideBanner": false
  },
  "tools": {
    "sandbox": false
  }
}

Tool Permissions & Rules

  • /permissions: add, edit, or remove allow/deny rules in-CLI
  • Rules merge across three layers: project, user, and CLI settings
  • Shared with the Antigravity desktop app's permission settings
  • Keep high-risk tools constrained for teams
# Manage permission rules interactively
> /permissions

# Auto-approve everything (use with care)
agy --dangerously-skip-permissions

📖 antigravity.google/docs

File Filtering

  • respectGitIgnore: Honor .gitignore patterns
  • enableRecursiveFileSearch: Recursive completion
  • Exclude rules and allowlists live in rules.json
{
  "context": {
    "fileFiltering": {
      "respectGitIgnore": true,
      "enableRecursiveFileSearch": true
    }
  }
}

Environment Variables

Variable Purpose
ANTIGRAVITY_API_KEY API-key authentication (alternative to sign-in)
AGY_CLI_DISABLE_LATEX Turn off LaTeX math rendering
AGY_CLI_HIDE_ACCOUNT_INFO Hide email and plan tier from the header
$EDITOR External editor for prompts and files
HTTP_PROXY Network proxy

Custom Context Filenames

  • AGENTS.md is the documented default context file
  • GEMINI.md and CLAUDE.md are also recognized
  • Support multiple filenames in priority order
  • Include additional directories for shared context
{
  "context": {
    "fileName": ["AGENTS.md", "GEMINI.md", "CLAUDE.md"],
    "includeDirectories": ["~/shared-context"],
    "loadMemoryFromIncludeDirectories": true
  }
}

layout: image-right image: https://images.unsplash.com/photo-1516321318423-f06f85e504b3?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Advanced Features

Power User Tools

MCP, Plugins, Sessions

Model Context Protocol (MCP)

  • Standard protocol for AI-to-system connections
  • Connect to external tools and services
  • Supports local commands, HTTP, and SSE
  • OAuth 2.0 for remote authentication

MCP Configuration: Firecrawl

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": ["@modelcontextprotocol/server-firecrawl"],
      "env": {
        "FIRECRAWL_API_KEY": "${FIRECRAWL_API_KEY}"
      }
    }
  }
}

MCP Configuration: Database

{
  "mcpServers": {
    "postgres": {
      "command": "npx",
      "args": ["@modelcontextprotocol/server-postgres"],
      "env": {
        "CONNECTION_STRING": "${DATABASE_URL}"
      }
    }
  }
}

Managing MCP Servers

Antigravity has no mcp subcommand — edit the config file directly:

# Open the MCP config
$EDITOR ~/.gemini/config/mcp_config.json
{
  "mcpServers": {
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"]
    }
  }
}

Servers initialize in parallel on startup.

MCP: Remote Servers

Use a url for HTTP/remote MCP servers (added in 1.0.5):

{
  "mcpServers": {
    "remote-tools": {
      "url": "https://mcp.example.com/sse"
    }
  }
}

Toggle servers from the in-CLI settings; changes reload on restart.

MCP Server Options

  • command: Shell command to start server
  • args: Command arguments
  • env: Environment variables
  • cwd: Working directory

MCP Server Options (Advanced)

  • timeout: Startup timeout in ms
  • includeTools/excludeTools: Filter available tools
  • trust: Bypass tool confirmation for this server

Popular MCP Servers

  • Firecrawl - Web scraping, search, content extraction
  • PostgreSQL - Database queries and schema
  • Filesystem - Extended file operations
  • Slack - Team communication
  • Playwright - Browser automation

📖 Registry: modelcontextprotocol.io/registry

Plugins from Marketplaces

  • Install from a marketplace with plugin@marketplace syntax
  • Plugins bundle skills and subagents, discovered automatically
  • Import existing Gemini or Claude plugins with agy plugin import
  • Plugins install to the shared ~/.gemini/config/ directory
# Install, then link a marketplace
agy plugin install my-plugin@my-marketplace
agy plugin link my-marketplace ./target

📖 antigravity.google/docs

Subagents

  • Dispatch focused tasks to a subagent with /agent <task>
  • Subagents run with their own context and interaction timeout
  • Bundled and discovered through installed plugins
  • Standalone subagent conversations stay separate in /resume
# Dispatch an async subagent
> /agent write unit tests for @./src/utils.py

Plugin System

  • Extend Antigravity CLI with plugins (skills + subagents)
  • Managed with the agy plugin subcommand
  • Import from Gemini or Claude: agy plugin import gemini
  • Installed plugins are scanned for skills and agents automatically
# List installed plugins
agy plugin list

# Install / enable / disable
agy plugin install my-plugin@marketplace
agy plugin enable my-plugin
agy plugin disable my-plugin

Skills

  • Package reusable expert workflows as skills
  • Distributed inside plugins, stored under ~/.gemini/config/skills/
  • Surface as slash commands via autocomplete
  • Discovered automatically from installed plugins
# Skills arrive with plugins
agy plugin install code-review@marketplace

# Invoke a skill-derived command in-session
> /code-review @./src

Session Management

  • Continue: pick up the most recent conversation
  • Resume by ID: jump to a specific past conversation
  • Browse: open the conversation picker with /resume
  • Conversations are stored in SQLite (.db)
# Continue the most recent conversation
agy -c

# Resume a specific conversation by ID
agy --conversation <id>

# Browse conversations from inside a session
> /resume

Non-Interactive (Print) Mode

  • -p / --print: run a single prompt and print the response
  • --print-timeout: bound how long print mode waits (default 5m)
  • Sandbox isolation is enforced in print mode too
  • Pipe input from other commands for scripting
# One-shot print mode
agy -p "List all TODO comments in @./src"

# Piped input
cat error.log | agy -p "Diagnose this stack trace"

Desktop Integration

  • Antigravity 2.0 desktop app: shares config and conversations
  • /export: push a CLI session into the desktop app
  • Shared permissions: rules carry across CLI and desktop
  • External editor: set $EDITOR for prompt and file editing
# Push this session to the desktop app
> /export

layout: image-left image: https://images.unsplash.com/photo-1498050108023-c5249f4df085?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Practical Applications

Real-World Workflows

Common use cases

Code Exploration

  • Understand unfamiliar codebases
  • Trace dependencies and data flow
  • Find patterns and conventions
  • Generate architecture documentation
agy -p "Analyze the architecture of @./src/ and explain
how the components interact"

agy -p "Trace the flow from the API endpoint to the database
in @./src/controllers/ and @./src/services/"

Test Generation

  • Generate unit tests for existing code
  • Identify edge cases automatically
  • Create integration test scaffolding
  • Mock setup and fixtures
agy -p "Create comprehensive unit tests for @./src/utils.py
with pytest, including edge cases"

agy -p "Generate integration tests for @./src/api/users.py
with proper mocking"

Documentation Generation

  • README files for projects
  • API documentation
  • Architecture diagrams (Mermaid)
  • Code comments and docstrings
agy -p "Generate a comprehensive README.md for this project"

agy -p "Add detailed docstrings to all public functions
in @./src/services/"

agy -p "Create a Mermaid diagram showing the system architecture"

Refactoring & Modernization

  • Upgrade legacy code patterns
  • Apply modern language features
  • Improve code organization
  • Fix anti-patterns
agy -p "Refactor @./src/legacy.py to use modern Python 3.12
features like type hints and match statements"

agy -p "Convert this callback-based code to async/await
@./src/api.js"

Debugging Workflows

  • Analyze error messages and stack traces
  • Identify root causes
  • Suggest fixes with context
  • Test and verify solutions
agy -p "This test is failing with @./tests/output.log.
Analyze the error and fix the issue in @./src/app.py"

agy -p "Debug why the API returns 500 errors.
Check @./src/routes.py and @./src/middleware.py"

Git Workflows

  • Generate commit messages
  • Create pull request descriptions
  • Analyze diffs and changes
  • Resolve merge conflicts
# Analyze staged changes
!git diff --staged
"Generate a conventional commit message for these changes"

# Create PR description
"Create a pull request description summarizing the changes
from the last 5 commits"

CI/CD Integration

  • Non-interactive --print mode for pipelines
  • Capture stdout to a file for parsing
  • Exit codes for success/failure
  • Automated code reviews
# In CI/CD pipeline
agy -p "Review @./src/ for security issues" > review.txt

# Check exit code
if agy -p "Verify all tests pass"; then
  echo "All checks passed"
fi

layout: image-right image: https://images.unsplash.com/photo-1522071820081-009f0129c71c?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Optional Advanced Module

  • Designed for teams moving from personal usage to production workflows
  • Can be taught live or assigned as follow-up practice
  • Focus areas: auth strategy, governance, skills/MCP, CI automation

Authentication Matrix (Team Use)

  • Google Sign-In: best default for local interactive usage
  • Device-code flow: shows a URL + code for remote/SSH machines
  • ANTIGRAVITY_API_KEY: scripts and headless automation
  • G1 credits: keep teams productive past standard quota

Auth Quick Checks

# Interactive route: sign in with Google on first run
agy

# API-key route (scripts / headless)
export ANTIGRAVITY_API_KEY="..."
agy -p "Summarize @./src"

# Hide account info from the header (e.g. for screen sharing)
export AGY_CLI_HIDE_ACCOUNT_INFO=1
agy

Governance: Permissions and Rules

  • --dangerously-skip-permissions is a per-session escape hatch
  • /permissions rules provide durable guardrails across users/projects
  • Keep high-risk tools constrained in team settings
  • Prefer explicit allow/deny patterns over ad-hoc approvals

Governance: Folder Trust

  • Antigravity tracks trusted folders (trustedFolders.json)
  • Folder trust impacts settings, skills, and context loading
  • Treat untrusted repos with --sandbox and stricter rules
  • Combine with /permissions for auditable guardrails

Skills + MCP at Scale

  • Skills package repeatable expert workflows for teams
  • MCP connects external systems (docs, data, browsers, APIs)
  • Distribute skills and subagents together inside plugins
  • Use includeTools/excludeTools to narrow risky MCP exposure

Plugins + MCP Commands

# Plugin (skills + subagents) lifecycle
agy plugin list
agy plugin disable my-plugin
agy plugin enable my-plugin

# MCP servers: edit config, then restart to reload
$EDITOR ~/.gemini/config/mcp_config.json

Automation Patterns (Non-Interactive)

  • Prefer --print for non-interactive runs
  • Use exit codes for pipeline gates
  • Keep prompts deterministic and repo-scoped
  • Store logs/artifacts for auditability

CI Example: Gate + Report

agy -p "Review @./src/ for security issues" > review.txt

if agy -p "Run checks and report pass/fail"; then
  echo "Gate passed"
else
  echo "Gate failed" && exit 1
fi

layout: image-right image: https://images.unsplash.com/photo-1522071820081-009f0129c71c?ixlib=rb-4.0.3&auto=format&fit=crop&w=1920&q=80 backgroundSize: cover

Best Practices

Professional Workflows

Tips for success

Effective Prompting

  • Be specific about what you want to achieve
  • Provide context about goals and constraints
  • Use file references to ground the conversation
  • Iterate for complex tasks
  • Include examples when possible

Safety Guidelines

  • Start in default mode - Get comfortable first
  • Enable checkpointing - Safety net for mistakes
  • Review generated code - Don't blindly accept
  • Use sandbox for exploratory work
  • Commit often - Git is your safety net

Project Setup

  1. Create a comprehensive AGENTS.md
  2. Set up project-specific settings.json
  3. Configure relevant MCP servers
  4. Create custom commands for common tasks
  5. Establish team conventions

AGENTS.md Best Practices

  • Keep it focused - Relevant project info only
  • Update regularly - Reflect current state
  • Use imports - Modularize large contexts
  • Include examples - Show expected patterns
  • Document conventions - Style guides, patterns

Team Collaboration

  • Share AGENTS.md in version control
  • Standardize settings.json across team
  • Create shared custom commands
  • Document AI-assisted workflows
  • Review AI-generated code together

Common Pitfalls

  • Overly broad prompts → Be specific
  • Missing context → Use AGENTS.md
  • Skipping review → Always verify output
  • Skipping permissions → Don't reach for --dangerously-skip-permissions early
  • Ignoring Artifact Review → Inspect diffs before applying

Troubleshooting

  • Authentication issues: Re-run agy to sign in, or check ANTIGRAVITY_API_KEY
  • Rate limits: Check /quota; G1 credits cover overflow
  • Tool failures: Verify mcp_config.json and server status
  • Context not loading: Confirm AGENTS.md location, reload session
  • Logs: Override the log path with --log-file
# Write logs to a known path for troubleshooting
agy --log-file ./agy.log

# Inspect loaded context and token usage
> /context

Quick Access

Antigravity CLI Docs

antigravity.google/docs

Course Repository

github.com/kousen/gemini-training

Important Links

📚 Official Documentation

https://antigravity.google/docs

💾 Source Repository

https://github.com/google-antigravity/antigravity-cli

📦 MCP Server Registry

https://modelcontextprotocol.io/registry

💻 Course Materials

https://github.com/kousen/gemini-training

Command Reference: Basic Usage

# Interactive mode
agy

# One-shot (print) mode
agy -p "prompt"

# Interactive with initial prompt
agy -i "context"

Command Reference: Safety

# Run in sandbox mode
agy --sandbox

# Auto-approve all tool calls (use with care)
agy --dangerously-skip-permissions

# Manage permission rules in-session
> /permissions

Command Reference: Sessions

# Continue the most recent conversation
agy -c

# Resume a conversation by ID
agy --conversation <id>

# Browse conversations in-session
> /resume

Command Reference: Output

# Print mode for scripting
agy -p "prompt"

# Bound how long print mode waits
agy --print-timeout 2m -p "prompt"

# Write logs to a file
agy --log-file ./agy.log

Thank You!

Questions?

Kenneth Kousen Author, Speaker, Java & AI Expert

kousenit.com | @kenkousen