API Reference

Complete reference for AgentOne commands, settings, and extensibility

🎯 Core Commands

Essential commands for basic AgentOne functionality.

AgentOne.plusButtonClicked

Command

Description: Opens the new task dialog to start a new coding task.

Keyboard Shortcut: None (available via UI button)

Usage: Click the "+" button in the AgentOne panel or run from Command Palette.

Command Palette Usage:

Ctrl+Shift+P (Cmd+Shift+P on Mac)
> AgentOne: New Task

AgentOne.addToChat

Command

Description: Adds selected text or current file to AgentOne chat context.

Keyboard Shortcut: Cmd+' (Mac) / Ctrl+' (Windows/Linux)

Context: Available when text is selected in the editor

Usage:

  1. Select text in any file
  2. Press Cmd+' or right-click → "Add to AgentOne"
  3. Selected text appears in chat context

AgentOne.openInNewTab

Command

Description: Opens AgentOne interface in a new VS Code tab for better multitasking.

Usage: Run from Command Palette when you need more screen space for AgentOne.

Benefits:

  • Larger chat interface
  • Side-by-side code editing
  • Multiple AgentOne sessions

AgentOne.focusChatInput

Command

Description: Focuses the chat input field for quick message typing.

Usage: Jump to chat input without using mouse, ideal for keyboard-driven workflows.

Workflow:

1. Working in editor
2. Ctrl+Shift+P → "AgentOne: Jump to Chat Input"
3. Type message immediately
4. Press Enter to send

🎼 Maestro Commands

Advanced commands for Maestro Mode functionality.

AgentOne.enableMaestroMode

Command

Description: Activates Maestro Mode for advanced multi-stage task execution.

Keyboard Shortcut: Cmd+Shift+M (Mac) / Ctrl+Shift+M (Windows/Linux)

Effect: Enables advanced reasoning, planning, and execution capabilities.

When to Use:

  • Complex refactoring tasks
  • Multi-file feature implementation
  • Architecture changes
  • Large-scale code generation

AgentOne.viewTaskPlan

Command

Description: Displays the current task execution plan with all phases and checkpoints.

Usage: Review how AgentOne plans to approach your task before execution begins.

Plan Information Includes:

  • Task breakdown and phases
  • File modification targets
  • Dependency analysis
  • Risk assessment
  • Rollback points

AgentOne.analyzeCodebase

Command

Description: Performs deep analysis of project structure, patterns, and conventions.

Context: Available when text is selected or run on entire project

Output: Comprehensive report on code architecture, patterns, and recommendations.

Analysis Includes:

  • Architectural patterns identification
  • Code quality metrics
  • Dependency analysis
  • Performance bottlenecks
  • Security considerations
  • Best practices compliance

🔧 Utility Commands

Helper commands for workflow optimization and interface management.

AgentOne.bestPracticesButtonClicked

Command

Description: Opens the Best Practices panel with automated development workflow tools.

Features: One-click access to code formatting, testing, security analysis, and documentation generation.

Available Best Practices:

Code Formatting & Style - Apply consistent formatting with Prettier, ESLint, Black, etc.
Automated Testing - Run test suites with coverage analysis
Security Analysis - Scan for vulnerabilities and security issues
Architecture Diagrams - Generate sequence and component diagrams
API Documentation - Create comprehensive API documentation
Dependency Management - Update and audit project dependencies

AgentOne.mcpButtonClicked

Command

Description: Opens the MCP (Model Context Protocol) servers management interface.

Purpose: Install, configure, and manage custom tools and integrations.

MCP Capabilities:

  • Browse community MCP servers
  • Install new tools and integrations
  • Configure server settings
  • Monitor server status
  • Create custom MCP servers

AgentOne.historyButtonClicked

Command

Description: Shows task execution history with ability to revert or repeat previous tasks.

Features: Browse past tasks, view execution details, and manage checkpoints.

History Features:

  • Task timeline with timestamps
  • File change summaries
  • Checkpoint management
  • Task repeat functionality
  • Export task reports

AgentOne.addTerminalOutputToChat

Command

Description: Adds terminal output to AgentOne chat for context and debugging.

Context: Available in terminal context menu

Usage: Right-click in terminal → "Add to AgentOne"

Common Use Cases:

  • Sharing error messages for debugging
  • Providing build output context
  • Including test results
  • Adding deployment logs

⚙️ Settings Reference

Complete reference for all AgentOne configuration options.

Core Settings

AgentOne.enableCheckpoints boolean

Default: true

Description: Enables Git-based checkpoints for rollback functionality. Requires Git to be initialized in the workspace.

Note: May not work well with very large repositories due to performance considerations.

AgentOne.disableBrowserTool boolean

Default: false

Description: Prevents AgentOne from spawning browser sessions for web automation tasks.

Use case: Enable in restricted environments or when browser automation isn't needed.

AgentOne.chromeExecutablePath string | null

Default: null

Description: Custom path to Chrome executable for browser automation. If not set, AgentOne will attempt to find or download Chrome automatically.

Example Paths:

// Windows
"C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe"

// macOS
"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"

// Linux
"/usr/bin/google-chrome"
AgentOne.preferredLanguage string

Default: "English"

Description: Language that AgentOne uses for communication and interface elements.

Options: English, Arabic, Portuguese (Brasil), Czech, French, German, Hindi, Hungarian, Italian, Japanese, Korean, Polish, Portuguese (Portugal), Russian, Simplified Chinese, Spanish, Traditional Chinese, Turkish

Maestro Mode Settings

AgentOne.maestroMode.enabled boolean

Default: false

Description: Enables the advanced Maestro agent mode for handling complex programming tasks.

AgentOne.maestroMode.analysisDepth "shallow" | "moderate" | "deep"

Default: "moderate"

Description: Controls the depth of analysis performed by the Maestro agent.

  • shallow: Quick analysis for simple tasks
  • moderate: Balanced analysis (recommended)
  • deep: Comprehensive analysis for complex projects
AgentOne.maestroMode.optimizationLevel "minimal" | "balanced" | "aggressive"

Default: "balanced"

Description: Controls how aggressively the Maestro agent optimizes code.

  • minimal: Focus on functionality over optimization
  • balanced: Good performance with maintainable code
  • aggressive: Maximum optimization for performance-critical applications
AgentOne.maestroMode.executionStrategy "speed" | "quality" | "efficiency"

Default: "quality"

Description: Prioritization strategy for executing complex tasks.

  • speed: Fastest implementation possible
  • quality: Focus on code quality and best practices
  • efficiency: Balance between speed and quality

Model-Specific Settings

AgentOne.modelSettings.o3Mini.reasoningEffort "low" | "medium" | "high"

Default: "medium"

Description: Controls the reasoning effort when using the o3-mini model. Higher values may result in more thorough but slower responses.

AgentOne.vsCodeLmModelSelector object

Description: Settings for VSCode Language Model API integration.

{
  "vendor": "copilot",
  "family": "gpt-4"
}

Integration Settings

AgentOne.mcpMarketplace.enabled boolean

Default: true

Description: Controls whether the MCP Marketplace is enabled for discovering and installing MCP servers.

📄 Configuration Files

AgentOne supports multiple configuration methods for different use cases.

Project Configuration (agentone.config.json)

Create this file in your project root for project-specific settings.

agentone.config.json
{
  "provider": "anthropic",
  "apiKey": "sk-ant-api03-...",
  "model": "claude-3.7-sonnet",
  "maxTokens": 4096,
  "temperature": 0.7,
  "maestroMode": {
    "enabled": true,  
    "analysisDepth": "moderate",
    "optimizationLevel": "balanced",
    "executionStrategy": "quality"
  },
  "enableCheckpoints": true,
  "disableBrowserTool": false,
  "preferredLanguage": "English"
}

VS Code Settings

Configure AgentOne through VS Code's settings interface or settings.json.

settings.json
{
  "AgentOne.enableCheckpoints": true,
  "AgentOne.maestroMode.enabled": true,
  "AgentOne.maestroMode.analysisDepth": "moderate",
  "AgentOne.preferredLanguage": "English",
  "AgentOne.chromeExecutablePath": "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
}

Environment Variables

Set API keys and sensitive configuration via environment variables.

Environment Variables
# Anthropic API Key
export ANTHROPIC_API_KEY="sk-ant-api03-..."

# OpenAI API Key  
export OPENAI_API_KEY="sk-..."

# Google API Key
export GOOGLE_API_KEY="AIza..."

# Custom model endpoint
export AGENTONE_BASE_URL="https://your-custom-endpoint.com"

Build with AgentOne

Use these APIs and configurations to customize AgentOne for your exact workflow needs.

Get Started View Source