# Auto-GPT: Recent Updates and New Features - May 2023

The past month has brought a bounty of updates and new features to the Auto-GPT repository. In this blog post, we will cover the most recent changes, highlighting additions to support environment variables, continuous integration improvements, testing enhancements, and exciting new challenges. With contributions from developers such as kaneda2004, merwanehamadi, Auto-GPT-Bot, Douglas Schonholtz, Erik Peterson, Luke K, Kinance, and Toran Bruce Richards, Auto-GPT continues to evolve and mature to provide an even more robust and user-friendly experience.

## Features

### Support for Environment Variables

On 31st May 2023, kaneda2004 made a significant commit that added support for the environment variable "OPENAI_ORGANIZATION" to the project's config. This modification allows users to specify their OpenAI organization key if applicable. The commit modified both the ".env.template" file and the "config.py" file. The pull request for this commit is available at [PR#289](https://github.com/Significant-Gravitas/Auto-GPT/pull/289).

## Continuous Integration

### Skipping Unnecessary Runs in CI

Merwanehamadi made a couple of relevant commits focused on improving the continuous integration process, helping to reduce CI runtime and save resources. The first commit, made on [PR#4482](https://github.com/Significant-Gravitas/Auto-GPT/pull/4482) added an exclusion for `current_score.json` in the integration test folder to prevent unnecessary CI runs based on changes to the `current_score` file. The second commit on [PR#4485](https://github.com/Significant-Gravitas/Auto-GPT/pull/4485) made updates to the GitHub Actions CI system to ignore the "tests/Auto-GPT-test-cassettes" directory and "tests/integration/challenges/current_score.json" file during the workflow, further optimizing the CI process.

## Testing

### VCR.py Integration and Test Improvements

Several commits throughout the month targeted test improvements and the implementation of VCR.py to simplify and streamline the process. Of note, merwanehamadi added the `@pytest.mark.vcr` decorator and made other adjustments related to VCR.py integration as part of [PR#4474](https://github.com/Significant-Gravitas/Auto-GPT/pull/4474). Erik Peterson, in [PR#4469](https://github.com/Significant-Gravitas/Auto-GPT/pull/4469), resolved issue #4461 by ensuring that error requests are not recorded during challenge testing.

## Challenges

### New Debug and Information Retrieval Challenges

The month of May brought exciting new challenges to the Auto-GPT project. Douglas Schonholtz created a debug challenge in commit #4286, which can be found at [this commit](https://github.com/Significant-Gravitas/Auto-GPT/commit/70be3c3e37091b76cfc68791cef5ca4e4dcd4c61). Additionally, merwanehamadi added a new challenge to the information retrieval section called "Challenge B," aiming to test the agent's ability to retrieve multiple pieces of related information in a consistent way. The commit link for this addition is available at [this commit](https://github.com/Significant-Gravitas/Auto-GPT/commit/bf77a984f64609c908389b5bc006164ec6f5fb1d).

## Conclusion

The Auto-GPT project has seen a remarkable selection of updates and improvements in May 2023, thanks to the hard work of our dedicated contributors. We're excited to see the ongoing development of this powerful AI tool and look forward to future achievements. Stay tuned for more updates, and be sure to visit [Auto-GPT on GitHub](https://github.com/Significant-Gravitas/Auto-GPT) to keep up with the latest developments!