### Build and Install CLI Locally

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Build the CLI tarball and then install it locally on your system. This allows you to test the installed CLI version.

```bash
# Build and install locally
./apps/cli/scripts/build.sh --install
```

--------------------------------

### IPC Client Usage Example

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/ipc/README.md

Demonstrates how to use the IpcClient to send commands to the Roo Code extension. This includes resuming an existing task and starting a new task with configuration and initial message.

```typescript
import { IpcClient } from "@roo-code/ipc"

const client = new IpcClient("/path/to/socket")

// Resume a task
client.sendCommand({
	commandName: "ResumeTask",
	data: "task-123",
})

// Start a new task
client.sendCommand({
	commandName: "StartNewTask",
	data: {
		configuration: {
			/* RooCode settings */
		},
		text: "Hello, world!",
		images: [],
		newTab: false,
	},
})
```

--------------------------------

### Install Project Dependencies

Source: https://github.com/zoo-code-org/zoo-code/blob/main/CONTRIBUTING.md

After cloning the repository, install all necessary project dependencies using pnpm. This command ensures your development environment is set up correctly.

```bash
pnpm install
```

--------------------------------

### Development Installation

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Installs dependencies and builds the Roo Code CLI from the monorepo for development purposes.

```bash
# From the monorepo root.
pnpm install

# Build the main extension first.
pnpm --filter ./src bundle

# Build the CLI.
pnpm --filter @roo-code/cli build
```

--------------------------------

### StartNewTask

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/ipc/README.md

Starts a new task with optional configuration and initial message. This command allows for the initiation of new tasks, providing initial data and settings.

```APIDOC
## StartNewTask

### Description
Starts a new task with optional configuration and initial message.

### Parameters
- **configuration** (object) - Required - RooCode settings object
- **text** (string) - Required - Initial task message
- **images** (array) - Optional - Array of image data URIs
- **newTab** (boolean) - Optional - Whether to open in a new tab
```

--------------------------------

### Quick Install Roo Code CLI

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Installs the Roo Code CLI using a curl command. Ensure Node.js 20+ is installed.

```bash
curl -fsSL https://raw.githubusercontent.com/RooCodeInc/Roo-Code/main/apps/cli/install.sh | sh
```

--------------------------------

### Custom Installation Directory

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Installs the Roo Code CLI to a custom directory using environment variables.

```bash
ROO_INSTALL_DIR=/opt/roo-code ROO_BIN_DIR=/usr/local/bin curl -fsSL ... | sh
```

--------------------------------

### Manual VSIX Installation via CLI

Source: https://github.com/zoo-code-org/zoo-code/blob/main/README.md

Install a manually built VSIX package using the VSCode command-line interface. Replace <version> with the actual version number.

```sh
code --install-extension bin/zoo-code-<version>.vsix
```

--------------------------------

### Automated VSIX Installation

Source: https://github.com/zoo-code-org/zoo-code/blob/main/README.md

Build and install the Zoo Code extension as a VSIX package automatically. This command handles uninstallation, building, and installation, prompting for editor selection by default.

```sh
pnpm install:vsix [-y] [--editor=<command>]
```

--------------------------------

### Example Debug Log Output

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

This is an example of the log output you might see when debug logging is enabled. It shows state transitions and message processing.

```text
[MessageProcessor] State update: {
  "messageCount": 5,
  "lastMessage": {
    "msgType": "ask:completion_result"
  },
  "stateTransition": "running → idle",
  "currentAsk": "completion_result",
  "isWaitingForInput": true
}
[MessageProcessor] EMIT waitingForInput: { "ask": "completion_result" }
[MessageProcessor] EMIT taskCompleted: { "success": true }
```

--------------------------------

### Install Specific Version

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Installs a specific version of the Roo Code CLI by setting the ROO_VERSION environment variable.

```bash
ROO_VERSION=0.1.0 curl -fsSL https://raw.githubusercontent.com/RooCodeInc/Roo-Code/main/apps/cli/install.sh | sh
```

--------------------------------

### Multi-turn Test Fixture Example

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

This JSON fixture demonstrates a two-turn test. The first turn matches a user message and responds with a tool call. The second turn matches the tool call ID and responds with an attempt completion.

```json
{
	"fixtures": [
		{
			"match": {
				"userMessage": "my-e2e-tag:my-test",
				"sequenceIndex": 0
			},
			"response": {
				"toolCalls": [{ "name": "read_file", "arguments": "{\"path\":\"marker.txt\"}", "id": "call_my_read" }]
			}
		},
		{
			"match": { "toolCallId": "call_my_read" },
			"response": {
				"toolCalls": [
					{ "name": "attempt_completion", "arguments": "{\"result\":\"MY_MARKER\"}", "id": "call_my_done" }
				]
			}
		}
	]
}
```

--------------------------------

### Build VSIX Package Manually

Source: https://github.com/zoo-code-org/zoo-code/blob/main/README.md

Generate the VSIX package for manual installation. This command creates a .vsix file in the bin/ directory.

```sh
pnpm vsix
```

--------------------------------

### Run Roo Code CLI in Stdin Stream Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Enables programmatic control by streaming NDJSON commands via stdin, used with --print. Supports specifying a taskId per start command.

```bash
printf '{"command":"start","requestId":"1","prompt":"1+1=?"}\n' | roo --print --stdin-prompt-stream --output-format stream-json
```

```bash
# Optional: provide taskId per start command
printf '{"command":"start","requestId":"1","taskId":"018f7fc8-7c96-7f7c-98aa-2ec4ff7f6d87","prompt":"1+1=?"}\n' | roo --print --stdin-prompt-stream --output-format stream-json
```

--------------------------------

### Uninstall Roo Code CLI

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Removes the Roo Code CLI installation files from the system.

```bash
rm -rf ~/.roo/cli ~/.local/bin/roo
```

--------------------------------

### Agent Response Messages

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

Send these messages to control the agent's behavior, such as approving/rejecting actions, answering questions, starting new tasks, or canceling current tasks.

```typescript
// Approve an action (tool, command, browser, MCP)
client.sendMessage({
	type: "askResponse",
	askResponse: "yesButtonClicked",
})
```

```typescript
// Reject an action
client.sendMessage({
	type: "askResponse",
	askResponse: "noButtonClicked",
})
```

```typescript
// Answer a question
client.sendMessage({
	type: "askResponse",
	askResponse: "messageResponse",
	text: "My answer here",
})
```

```typescript
// Start a new task
client.sendMessage({
	type: "newTask",
	text: "Build a web app",
})
```

```typescript
// Cancel current task
client.sendMessage({
	type: "cancelTask",
})
```

--------------------------------

### Run CLI Directly from Source

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Execute the CLI using pnpm dev, specifying the provider and API key. Useful for quick development cycles without a full build.

```bash
# Run directly from source (no build required)
pnpm dev --provider openrouter --api-key $OPENROUTER_API_KEY --print "Hello"
```

--------------------------------

### Initialize and Use ExtensionClient

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

Instantiate the ExtensionClient to manage agent state and communicate with the extension. You can query the agent's state and current mode, and subscribe to state and mode change events.

```typescript
const client = new ExtensionClient({
	ssendMessage: (msg) => extensionHost.sendToExtension(msg),
	debug: true, // Writes to ~/.roo/cli-debug.log
})

// Query state at any time
const state = client.getAgentState()
if (state.isWaitingForInput) {
	console.log(`Agent needs: ${state.currentAsk}`)
}

// Query current mode
const mode = client.getCurrentMode()
console.log(`Current mode: ${mode}`) // e.g., "code", "architect", "ask"

// Subscribe to events
client.on("waitingForInput", (event) => {
	console.log(`Waiting for: ${event.ask}`)
})

// Subscribe to mode changes
client.on("modeChanged", (event) => {
	console.log(`Mode changed: ${event.previousMode} -> ${event.currentMode}`)
})
```

--------------------------------

### Verify xAI Grok Fixtures in Mock Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Run this command to verify recorded xAI fixtures without hitting the real API. No API key is needed. This uses the `fixtures/xai.json` for responses.

```sh
TEST_FILE=xai.test pnpm --filter @roo-code/vscode-e2e test:ci:mock
```

--------------------------------

### Create a Named Fixture File

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Define a named fixture file with a short, stable match string and the expected LLM response. Ensure the match string does not contain timestamps or environment details.

```json
{
	"fixtures": [
		{
			"match": { "userMessage": "your unique prompt text" },
			"response": {
				"toolCalls": [
					{ "name": "attempt_completion", "arguments": "{\"result\":\"...\"}", "id": "call_001" }
				]
			}
		}
	]
}
```

--------------------------------

### Fast Local CLI Build

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Perform a quick build of the CLI tarball, skipping the verification step. Use this for faster iteration during development when verification is not critical.

```bash
# Fast build (skip verification)
./apps/cli/scripts/build.sh --skip-verify
```

--------------------------------

### Clone the Zoo-Code Repository

Source: https://github.com/zoo-code-org/zoo-code/blob/main/CONTRIBUTING.md

Clone the main repository to your local machine to begin development. Ensure you replace YOUR_USERNAME with your GitHub username.

```bash
git clone https://github.com/YOUR_USERNAME/Zoo-Code.git
```

--------------------------------

### Override Backend URL for Local Development

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Run the CLI in development mode while pointing the backend URL to production. This is useful for testing against the live API.

```bash
ROO_CODE_PROVIDER_URL=https://api.roocode.com/proxy pnpm dev --provider openrouter --api-key $OPENROUTER_API_KEY --print "Hello"
```

--------------------------------

### Configure Extension Dependencies

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/types/npm/README.md

Add the Zoo Code extension to your extension's `extensionDependencies` in `package.json` to ensure it is activated before your extension.

```json
"extensionDependencies": ["ZooCodeOrganization.zoo-code"]
```

--------------------------------

### Run CLI Tests

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Execute the test suite for the CLI application using pnpm test.

```bash
# Run tests
pnpm test
```

--------------------------------

### Verify Tests in Mock Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Run E2E tests in mock mode without needing an API key to verify that fixtures are correctly set up and replayed.

```sh
pnpm --filter @roo-code/vscode-e2e test:ci:mock
```

--------------------------------

### Verify DeepSeek V4 Fixtures in Mock Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Verify DeepSeek V4 fixtures using aimock. Set `USE_MOCK=true` to ensure responses come from fixtures. No API key is needed for this verification step.

```sh
USE_MOCK=true TEST_FILE=deepseek-v4.test pnpm --filter @roo-code/vscode-e2e test:run
```

--------------------------------

### Record xAI Grok Fixtures

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Use this command to record SSE events from the real xAI API. Ensure you have an XAI_API_KEY set. This populates the gitignored `fixtures/xai.json` file.

```sh
XAI_API_KEY=<key> XAI_RECORD=true TEST_FILE=providers/xai.test.js pnpm --filter @roo-code/vscode-e2e test:run
```

--------------------------------

### CLI Architecture Diagram

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

A visual representation of the Zoo CLI's internal architecture, showing the flow from the CLI entry point to the ExtensionHost and its interaction with VSCode and the Extension Bundle.

```text
┌─────────────────┐
│   CLI Entry     │
│   (index.ts)    │
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  ExtensionHost  │
│  (extension-    │
│   host.ts)      │
└────────┬────────┘
         │
    ┌────┴────┐
    │         │
    ▼         ▼
┌───────┐  ┌──────────┐
│vscode │  │Extension │
│-shim  │  │ Bundle   │
└───────┘  └──────────┘
```

--------------------------------

### Record Fixtures with OpenRouter

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Record E2E test fixtures using an OpenRouter API key. This command filters the recording to a specific test file to avoid re-recording unrelated tests.

```sh
OPENROUTER_API_KEY=<key> pnpm --filter @roo-code/vscode-e2e test:record
```

--------------------------------

### Lint CLI Code

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Apply linting rules to the CLI codebase using pnpm lint.

```bash
# Linting
pnpm lint
```

--------------------------------

### Record Fixtures for a Specific Test File

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Record fixtures for a single test file by specifying the TEST_FILE environment variable. This helps to isolate recording to the relevant tests.

```sh
OPENROUTER_API_KEY=<key> TEST_FILE=my-feature.test.js pnpm --filter @roo-code/vscode-e2e test:record
```

--------------------------------

### CloseTask

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/ipc/README.md

Closes a task and performs cleanup. This command is used to terminate a task and execute any necessary cleanup procedures.

```APIDOC
## CloseTask

### Description
Closes a task and performs cleanup.

### Parameters
- **data** (string) - Required - Task ID to close
```

--------------------------------

### Record DeepSeek V4 Fixtures

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Record fixtures for DeepSeek V4 by proxying OpenAI-compatible traffic through aimock. Ensure you have a DEEPSEEK_API_KEY set. This generates `openai-*.json` files that should be converted to stable named fixtures.

```sh
DEEPSEEK_API_KEY=<key> TEST_FILE=deepseek-v4.test pnpm --filter @roo-code/vscode-e2e test:record
```

--------------------------------

### Run Roo Code CLI in Print Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Executes the Roo Code CLI for non-interactive, machine-readable output. A prompt is required.

```bash
# Prompt is required
roo --print "Summarize this repository"
```

```bash
# Create a new task with a specific session ID (UUID)
roo --print --create-with-session-id 018f7fc8-7c96-7f7c-98aa-2ec4ff7f6d87 "Summarize this repository"
```

--------------------------------

### Record Fixtures with Anthropic Provider

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Record fixtures for tests that use the Anthropic provider directly. Both OpenRouter and Anthropic API keys are required, and the test file can be specified.

```sh
OPENROUTER_API_KEY=<or-key> ANTHROPIC_API_KEY=<key> TEST_FILE=my-anthropic-test.test.js pnpm --filter @roo-code/vscode-e2e test:record
```

--------------------------------

### Gemini Mock Mode Test Command

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Command to verify Gemini tests in mock mode. This does not require a key and uses the recorded fixtures.

```sh
TEST_FILE=providers/gemini.test pnpm --filter @roo-code/vscode-e2e test:ci:mock
```

--------------------------------

### Z.ai GLM Test Commands

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Commands to run Z.ai GLM tests. The default mode uses a fetch interceptor. To bypass the interceptor and hit the real API, set the ZAI_API_KEY environment variable.

```sh
# Mock mode (default — no key needed, interceptor active)
pnpm --filter @roo-code/vscode-e2e test:ci:mock
```

```sh
# Live mode — bypasses interceptor, calls real Z.ai API
ZAI_API_KEY=<key> TEST_FILE=zai.test pnpm --filter @roo-code/vscode-e2e test:ci
```

--------------------------------

### Build CLI Tarball for Current Platform

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Generate a platform-specific tarball of the CLI application using the local build script. This is useful for testing builds outside of the CI environment.

```bash
# Build tarball for your current platform
./apps/cli/scripts/build.sh
```

--------------------------------

### Find Auto-Recorded Fixture File

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Search for the auto-recorded fixture file corresponding to your test by using grep with a unique text string from your prompt.

```sh
grep -l "your unique prompt text" apps/vscode-e2e/fixtures/openai-*.json
```

--------------------------------

### Clean Stale Fixture Files

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Use this command to remove any old auto-recorded fixture files before recording new ones. The -x flag is necessary to clean gitignored files.

```sh
git clean -fx apps/vscode-e2e/fixtures/
```

--------------------------------

### Configure Non-Default Anthropic Provider

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Use this configuration when your tests call `api.setConfiguration` with `apiProvider: "anthropic"`. Ensure `anthropicBaseUrl` is provided to point aimock at the Anthropic endpoint. Remember to restore the default OpenRouter config in `suiteTeardown`.

```typescript
await api.setConfiguration({
	apiProvider: "anthropic" as const,
	apiKey: aimockUrl && !isRecord ? "mock-key" : process.env.ANTHROPIC_API_KEY!,
	apiModelId: "claude-opus-4-7",
	...(aimockUrl && { anthropicBaseUrl: aimockUrl }),
})
```

--------------------------------

### Run Roo Code CLI in Approval-Required Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Executes the Roo Code CLI requiring manual approval for tools, commands, browser, and MCP actions.

```bash
roo "Refactor the utils.ts file" --require-approval -w ~/Documents/my-project
```

--------------------------------

### Clone Zoo Code Repository

Source: https://github.com/zoo-code-org/zoo-code/blob/main/README.md

Use this command to clone the Zoo Code repository from GitHub to your local machine.

```sh
git clone https://github.com/Zoo-Code-Org/Zoo-Code.git
```

--------------------------------

### Enable and View Debug Logging

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

Enable debug logging for the CLI using the '-d' flag. Logs are stored in '~/.roo/cli-debug.log' and can be monitored in real-time.

```bash
roo -d -P "Build something" --no-tui
```

```bash
tail -f ~/.roo/cli-debug.log
```

--------------------------------

### Perform CLI Type Checking

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Check the TypeScript types of the CLI project using pnpm check-types.

```bash
# Type checking
pnpm check-types
```

--------------------------------

### Gemini Record Mode Test Command

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Command to record fixtures for Gemini tests. This requires a GEMINI_API_KEY and will proxy to the real Gemini API. After recording, inspect and extract response blocks into fixtures/gemini.json.

```sh
GEMINI_API_KEY=<key> TEST_FILE=providers/gemini.test pnpm --filter @roo-code/vscode-e2e test:record
```

--------------------------------

### Add Programmatic Fixture with RegExp Match

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

Add programmatic fixtures in `src/runTest.ts` using `mock.addFixture()` with a `RegExp` match for requests that cannot be matched by a stable substring. These fixtures are only available in replay mode and are not recorded.

```typescript
mock.addFixture()
```

--------------------------------

### Run Roo Code CLI in Interactive Mode

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Executes the Roo Code CLI in interactive TUI mode with auto-approval for actions and commands. Requires OPENROUTER_API_KEY to be set.

```bash
export OPENROUTER_API_KEY=sk-or-v1-...

roo "What is this project?" -w ~/Documents/my-project
```

```bash
roo -w ~/Documents/my-project
```

--------------------------------

### API Methods

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/types/npm/README.md

This section details the methods available in the Zoo Code API for interacting with the extension.

```APIDOC
## API Methods

### Description
This section details the methods available in the Zoo Code API for interacting with the extension.

### Methods

- **startNewTask**(message: string, images?: string[]): Promise<void>
  - Description: Starts a new task with an optional initial message and a list of images.
  - Parameters:
    - `message` (string) - Required - The initial message for the new task.
    - `images` (string[]) - Optional - A list of image URIs to include with the task.

- **sendMessage**(message: string): Promise<void>
  - Description: Sends a message to the current task.
  - Parameters:
    - `message` (string) - Required - The message to send.

- **pressPrimaryButton**(): Promise<void>
  - Description: Simulates pressing the primary button in the chat interface (e.g., 'Save' or 'Proceed While Running').

- **pressSecondaryButton**(): Promise<void>
  - Description: Simulates pressing the secondary button in the chat interface (e.g., 'Reject').
```

--------------------------------

### VS Code E2E Agent Test Commands

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/vscode-e2e/AGENTS.md

These commands are used to run end-to-end tests for the VS Code agent in different modes: replay, record, and live API.

```sh
pnpm --filter @roo-code/vscode-e2e test:ci:mock
```

```sh
OPENROUTER_API_KEY=<key> pnpm --filter @roo-code/vscode-e2e test:record
```

```sh
OPENROUTER_API_KEY=<key> pnpm --filter @roo-code/vscode-e2e test:ci
```

--------------------------------

### StateStore Interface Definition

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

Defines the structure for storing the agent's state, including messages, computed state information, initialization status, and the current mode.

```typescript
interface StoreState {
	messages: ClineMessage[] // The raw message array
	agentState: AgentStateInfo // Computed state
	isInitialized: boolean // Have we received any state?
	currentMode: string | undefined // Current mode (e.g., "code", "architect")
}
```

--------------------------------

### ResumeTask

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/ipc/README.md

Resumes a task from history. This command allows for the continuation of a previously executed task by referencing its ID.

```APIDOC
## ResumeTask

### Description
Resumes a task from history.

### Parameters
- **data** (string) - Required - Task ID to resume
```

--------------------------------

### Import and Use Zoo Code API

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/types/npm/README.md

Import the RooCodeAPI type and load the Zoo Code extension to access its functionalities. Ensure the extension is active before using the API.

```typescript
import { RooCodeAPI } from "@roo-code/types"

const extension = vscode.extensions.getExtension<RooCodeAPI>("ZooCodeOrganization.zoo-code")

if (!extension?.isActive) {
	throw new Error("Extension is not activated")
}

const api = extension.exports

if (!api) {
	throw new Error("API is not available")
}

// Start a new task with an initial message.
await api.startNewTask("Hello, Zoo Code API! Let's make a new project...")

// Start a new task with an initial message and images.
await api.startNewTask("Use this design language", ["data:image/webp;base64,..."])

// Send a message to the current task.
await api.sendMessage("Can you fix the @problems?")

// Simulate pressing the primary button in the chat interface (e.g. 'Save' or 'Proceed While Running').
await api.pressPrimaryButton()

// Simulate pressing the secondary button in the chat interface (e.g. 'Reject').
await api.pressSecondaryButton()
```

--------------------------------

### Upgrade Roo Code CLI

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Upgrades the Roo Code CLI using the built-in upgrade command.

```bash
roo upgrade
```

--------------------------------

### Check Legacy Roo Auth Token Status

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Checks if a legacy Roo auth token is stored on the system.

```bash
roo auth status
```

--------------------------------

### Type Guards for Message Categorization

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

Utilize type guards from '@roo-code/types' to categorize incoming agent messages. This helps in determining the state and required action for the agent.

```typescript
import { isIdleAsk, isInteractiveAsk, isResumableAsk, isNonBlockingAsk } from "@roo-code/types"

const ask = message.ask
if (isInteractiveAsk(ask)) {
	// Needs approval: tool, command, followup, etc.
} else if (isIdleAsk(ask)) {
	// Task stopped: completion_result, api_req_failed, etc.
} else if (isResumableAsk(ask)) {
	// Task paused: resume_task
} else if (isNonBlockingAsk(ask)) {
	// Command running: command_output
}
```

--------------------------------

### Streaming Detection Logic

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

This pseudocode function determines if the agent is currently streaming a response. It checks for a `partial: true` flag on the last message or an incomplete API request indicated by `cost: undefined`.

```typescript
function isStreaming(messages) {
	const lastMessage = messages.at(-1)

	// Check partial flag (primary indicator)
	if (lastMessage?.partial === true) {
		return true
	}

	// Check for in-progress API request
	const apiReq = messages.findLast((m) => m.say === "api_req_started")
	if (apiReq?.text) {
		const data = JSON.parse(apiReq.text)
		if (data.cost === undefined) {
			return true // API request not yet complete
		}
	}

	return false
}
```

--------------------------------

### ClineMessage Interface Definition

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/docs/AGENT_LOOP.md

Defines the structure of messages exchanged between the agent and the client, including timestamp, type, and content details.

```typescript
interface ClineMessage {
	ts: number // Unique timestamp identifier
	type: "ask" | "say" // Message category
	ask?: ClineAsk // Specific ask type (when type="ask")
	say?: ClineSay // Specific say type (when type="say")
	text?: string // Message content
	partial?: boolean // Is this message still streaming?
}
```

--------------------------------

### Remove Legacy Roo Auth Token

Source: https://github.com/zoo-code-org/zoo-code/blob/main/apps/cli/README.md

Logs out and removes any old stored Roo auth token from the system.

```bash
roo auth logout
```

--------------------------------

### CancelTask

Source: https://github.com/zoo-code-org/zoo-code/blob/main/packages/ipc/README.md

Cancels a running task. This command is used to stop an ongoing task by providing its unique identifier.

```APIDOC
## CancelTask

### Description
Cancels a running task.

### Parameters
- **data** (string) - Required - Task ID to cancel
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.