<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Updated the Browserbase session creation to enable proxies and remove
advanced stealth settings for improved evaluation behavior.
<!-- End of auto-generated description by cubic. -->
Better docs icon, same as in cloud
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Updated the docs favicon to match the cloud version for a consistent
look.
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Changed the screenshot test to start browser sessions one at a time
instead of in parallel to avoid Playwright semaphore errors.
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Improved error handling for missing optional imports in eval by raising
clear ImportError messages with install instructions.
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Added support for Browserbase and Hyperbrowser as remote browser options
in the eval service.
- **New Features**
- Users can now select Browserbase or Hyperbrowser with the `--browser`
flag.
- Added environment variable checks and session creation logic for both
services.
<!-- End of auto-generated description by cubic. -->
refactor: simplify email query parameter description and enhance action
result messages
- Updated the description of the `query` parameter in
`GetRecentEmailsParams` for clarity.
- Refined the action result messages for better readability and
consistency, including specific query details in the long-term memory
response.
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Simplified the Gmail email query parameter description and improved
action result messages for clearer feedback.
- **Refactors**
- Shortened and clarified the `query` parameter description.
- Updated action result messages to include query details and improve
readability.
<!-- End of auto-generated description by cubic. -->
Improved filesystem error handling and added better validation for
filenames
Updated model names everywhere to gpt-4.1
Added an example of writing into csv
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
The eval service now fetches rotating authentication info from the
server and injects it into tasks that require login credentials.
- **New Features**
- Added server call to fetch auth distribution if any tasks specify
required auth keys.
- Injects relevant login credentials into task descriptions for agent
use.
<!-- End of auto-generated description by cubic. -->
- Added new file types: JsonFile and CsvFile to the file system.
- Updated read_file method to handle external files with .json and .csv extensions.
- Modified write_file action to allow .json and .csv file extensions in addition to .md and .txt.
- Introduced pypdf dependency for PDF file handling.
This update improves the flexibility of file operations within the application.
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Added support for selecting different browsers (local, Anchor Browser,
Brightdata) in the eval script using a new --browser argument.
- **New Features**
- Users can choose the browser backend with --browser (local,
anchor-browser, brightdata, browser-use).
- Updated environment variable handling and logging for new browser
options.
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Added support for using Anchor Browser when running evals, allowing
remote browser sessions if an API key is set.
- **New Features**
- Detects ANCHOR_BROWSER_API_KEY and connects to Anchor Browser
automatically.
- Falls back to local browser if Anchor Browser is unavailable.
<!-- End of auto-generated description by cubic. -->
