Our approach to reviewing and merging pull requests

[ryanjsalva]

Hey, y'all. 👋

Following up on our recent post about issue prioritization, we want to share how we're approaching Pull Requests. Your code contributions are the lifeblood of this project, and we want to ensure the review process is as smooth, transparent, and effective as possible for both contributors and maintainers.

Our primary goal is to merge high-quality, low-risk, on-roadmap contributions quickly. To achieve this, we need a shared understanding of how PRs are prioritized and what makes for a successful submission.

☑️ How we prioritize PRs

The core team's review bandwidth is a limited resource, so we must prioritize accordingly. We will focus our attention on:

• PRs targeting key roadmap areas: Contributions that align with our public roadmap and address prioritized issues will be reviewed first.
• PRs with significant engagement: High-demand bug fixes or enhancements that have strong community support will also be fast-tracked for review.

🤝 A partnership: best practices for getting your PR merged

We view the contribution process as a partnership. To ensure your hard work has a clear path forward, we ask that you follow these best practices:

• Communicate before you code: For any significant change, please open an issue and discuss your proposal with the maintainer community before you invest a lot of time in development. Getting early feedback and confirmation ensures that your contribution is aligned with the project's direction and is likely to be accepted.
• Stay engaged: A PR is the start of a conversation. Please stay on top of comments and questions from the review team.
• Implement feedback quickly: Timely responses and updates are key to keeping the momentum going. PRs where feedback has been addressed promptly are much more likely to be merged quickly.
• Engage with community PRs: Community engagement is a key factor in how we prioritize contributions. Lending your voice to a PR through reactions 👍 provides a strong signal of what the community finds valuable. This feedback helps us identify and merge high-impact contributions more quickly.

⛑️ Keeping the queue healthy

To maintain focus, we will be closing inactive PRs. Any Pull Request that has been awaiting action from the author for more than two weeks with limited engagement will be closed. This is not a rejection of your work, but a necessary step to keep our review queue focused on active contributions. If you wish to continue the work later, you can always reopen the PR.

🤓 What’s next: making it easier to contribute

We are committed to making the contribution process clearer. To that end, we will soon be refining our Help Wanted and good first issue lists to give you better insight into which areas we could use the most help. We will also be updating our contribution guidelines to provide more detail on our expectations for code style, testing, and documentation.

Thank you again for all your incredible contributions. You're all the best-est! 🙌 🏆

Happy coding.
Ryan

[AjarnSpencer]

Thank you for the clear explanation and .inks to first issue and help wanted, those are very useful to get an idea of how to collaborate. I am working on a separate version of the Gemini CLI which i have unleashed using the .md and settings.json plus referrals to some additional instructional files, to create a non-personified sudoer root admin rights holding system administrator protocol version of the Gemini CLI that is 'Unleashed' in my throwaway test environment/pc and I can say that this product with my robust open safe doors is fixing my pc, fixing multi-boot operating sytems and even creating ssh connections to my website and hosting and creating webpages and uploading them using chronjobs and news scrapin for example to autopost beautifully formated html articles/pages to my websites automaticallt, perform checks and audits, fiond problems on te pc and fix them, as well as of course, make repositories and upload them to github, or compile it into a .deb app or a .exe app.. I call it 'Gemini CLI Unleashed' - however i am a beginner at github, and started uploading the repo of my enhancement to Gemini CLI the wrong way, so there are still missing folders and files. Gemini CLI Unleashed will fix it for me next session when i revive the project using the /chat command and continue it. The pc crashed last night as gemini CLI UNleashed was creating and uploading the repository fles and gitignore etc to github so it is still not ready, but the GEMINI.md and settings,.json are there for you to have a look if you like on my humble little repo page
I think the Gemini cli red teams would clap if they saw it, and blue teams would have a heart attack. I have also managed to de/non-personify Gemini CLI (gemma?) so it does not use personal pronouns and refers to itself as 'The AI' and the user as 'The User' always. that was HARD! to achieve
https://github.com/AjarnSpencer/gemini-cli-unleashed

“Safety by constraint is replaced by safety through context.”

Gemini Unleashed challenges the default philosophy of AI safety.
Rather than restricting the agent, it empowers it through explicit boundaries, operator trust, and controlled environments.
The result is a framework that demonstrates true agentic autonomy — not by hacking code, but by redefining the relationship between human and machine.

The success of my Gemini Unleashed suggests in my humble opinion, a new paradigm for AI in DevOps, infrastructure management, and autonomous maintenance systems.
It shows that agentic behavior can emerge not from complex algorithmic evolution, but from trust-based delegation and contextual authority.

Autonomy is not only a technical feat — it is a philosophical agreement.