### Installing AgentScript via npm

Source: https://github.com/agentscript-ai/agentscript/blob/main/README.md

This command installs the AgentScript SDK as a project dependency using the Node Package Manager (npm).

```shell
npm install agentscript-ai
```

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

### Generated AgentScript Code

Source: https://github.com/agentscript-ai/agentscript/blob/main/README.md

This JavaScript code is an example of code generated by an LLM using AgentScript to find and summarize issues. The generated code calculates a date one week in the past, searches for issues created after that date, and then summarizes the retrieved issues to provide a progress update. The `result` variable stores the outcome of the agent.

```javascript
// NOTE: this is real code generated by LLM.
// it's not executed but parsed into AST and runs in a specialized runtime.

// Calculate the date from 7 days ago
const lastWeek = addToDate(new Date(), { days: -7 });

// Search for all issues created in the last week
const issues = linear.searchIssues({
    createdAfter: lastWeek,
    orderBy: 'createdAt',
});

// Create a summary of the issues found
result = summarizeData({
    data: issues,
    prompt: 'Create a progress update focusing on number of tasks created, their current status distribution, and any notable patterns. Format as a clear business update.',
});
```

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

### Executing the Agent and Logging Output

Source: https://github.com/agentscript-ai/agentscript/blob/main/README.md

This TypeScript code shows how to execute an AgentScript agent and log its output and execution state. It calls `executeAgent` to run the agent, then logs the agent's output and the variables in the root execution context to the console.

```typescript
await executeAgent({ agent });

// See the output
console.log(agent.output);

// Check variables in the execution state
console.log(agent.root.variables);
```

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

### Defining an Agent with TypeScript

Source: https://github.com/agentscript-ai/agentscript/blob/main/README.md

This TypeScript code defines and executes an AI agent using the AgentScript SDK. It configures a language model, a Linear client, defines available tools including `addToDate`, `summarizeData`, and `searchIssues`, sets a task prompt, specifies the expected output type, and then generates and executes the agent.

```typescript
import { executeAgent, inferAgent } from 'agentscript-ai';
import * as s from 'agentscript-ai/schema';
import { addToDate, summarizeData } from 'agentscript-ai/tools';
import { LinearClient, searchIssues } from '@agentscript-ai/linear';
import { anthropic } from '@ai-sdk/anthropic';

// Configure the language model
const model = anthropic('claude-3-5-sonnet-latest');

// Configure the Linear client
const linear = LinearClient({
    apiKey: process.env.LINEAR_API_KEY,
});

// Define available tools.
const tools = {
    // Needed for date calculation
    addToDate,
    // Turns data into text
    summarizeData: summarizeData({ model }),
    // The real deal
    linear: {
        searchIssues: searchIssues({ model, linear }),
    },
};

// Define a task for the agent
const prompt = 'Give me a progress update of tasks created in the last week';

// Define the expected output
const output = s.string();

// Let the LLM generate the AgentScript code
const agent = await inferAgent({
    tools,
    output,
    model,
    prompt,
});
```

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

### AgentScript Type Definitions

Source: https://github.com/agentscript-ai/agentscript/blob/main/README.md

This TypeScript code provides type definitions for the AgentScript environment, including the `Duration` type for time intervals, function signatures for `addToDate` and `summarizeData`, and the `linear` namespace with types for `Issue` and `searchIssues`.  It also declares the global `result` variable where the agent's final output should be stored.

```typescript
export type Duration = {
    years?: number;
    months?: number;
    days?: number;
    hours?: number;
    minutes?: number;
    seconds?: number;
};

/** Add a duration to a date. */
export function addToDate(date: Date, duration: Duration): Date;

export type SummarizeDataParams = {
    /** The data to summarize. Can be in any format. */
    data: unknown;
    /**
     * The prompt to use to summarize the data.
     * Describe the expected outcome.
     */
    prompt: string;
};

/** Summarize any data */
export function summarizeData(params: SummarizeDataParams): string;

declare namespace linear {
    export type Issue = {
        id: string;
        url: string;
        title: string;
        description?: string;
        status: string;
        createdAt: Date;
        updatedAt: Date;
    };

    /**
     * Search for issues using a natural language query.
     * Do not filter results later, put all the search criteria in the query.
     * @param query - Descriptive query in object format.
     */
    export function searchIssues(query: unknown): Issue[];
}

/** You must put the result of the task here. */
let result: string;
```

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