chore: version packages

[openrouter-port-bot]

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@openrouter/agent@1.0.0

Major Changes

• #110 9766f31 Thanks @LukasParke! - Stop accumulating generator-tool preliminaryResults arrays. Yields are still broadcast live; the terminal tool.result event and ToolExecutionResult no longer carry the full yield history.

Minor Changes

• #118 4dce84e Thanks @sambarnes! - Allow manual tools to provide a caller-owned JSON Schema for wire serialization.

  import { tool } from "@openrouter/agent";
  import { z } from "zod";
  
  const confirmTool = tool({
    name: "confirm_action",
    inputSchema: z.object({ action: z.string() }),
    wireInputSchema: {
      type: "object",
      properties: {
        action: { type: "string" },
      },
      required: ["action"],
    },
    execute: false,
  });

• #93 dacf4d1 Thanks @LukasParke! - Fix fromChatMessages dropping assistant tool calls and give both message converters precise array return types that work with callModel, Item[], and the SDK's InputsUnion.

  Assistant toolCalls now become function_call items, preserving their already-serialized arguments. A message containing both text and tool calls emits both items; an empty assistant message is omitted only when tool calls replace it.

  import {
    callModel,
    fromChatMessages,
    type ChatMessages,
    type Item,
  } from "@openrouter/agent";
  
  const messages: ChatMessages[] = [
    {
      role: "assistant",
      content: null,
      toolCalls: [
        {
          id: "call_1",
          type: "function",
          function: { name: "get_weather", arguments: '{"city":"Austin"}' },
        },
      ],
    },
  ];
  
  const input: Item[] = fromChatMessages(messages);
  const result = callModel(client, { model: "openai/gpt-4o-mini", input });

Patch Changes

• #120 ddab365 Thanks @LukasParke! - Identify Agent SDK requests with an Agent SDK user-agent suffix that includes the package version.

  import { OpenRouter } from "@openrouter/agent";
  
  const client = new OpenRouter({ apiKey: process.env.OPENROUTER_API_KEY });
  // Requests use the Agent SDK user agent by default.
  // Pass userAgent to override the default identification.

• #125 ddd60df Thanks @w0nche0l! - Do not execute tool calls from a response truncated at max_output_tokens. A turn that hits the token budget mid-tool-call now finalizes with status: 'incomplete' instead of running the cut-off call (and re-requesting on the same exhausted budget).

  import { callModel } from "@openrouter/agent";
  
  const result = callModel(client, {
    model: "anthropic/claude-sonnet-4.5",
    maxOutputTokens: 512, // budget exhausted mid tool call
    tools: [myTool],
  });
  const response = await result.getResponse();
  // response.status === 'incomplete'
  // response.incompleteDetails.reason === 'max_output_tokens'
  // The cut-off call is NOT executed; raise the budget and re-request.

@openrouter/mcp@1.1.2

Patch Changes

• Updated dependencies [4dce84e, 9766f31, dacf4d1, ddab365, ddd60df]:
  • @openrouter/agent@1.0.0