[Showcase/Feedback] Steering Studio - starter context generation for GitHub Copilot Agent Mode and Kiro

[ejochens]

🏷️ Discussion Type

Product Feedback

💬 Feature/Topic Area

Copilot in GitHub

Body

I wanted to share a project I built that may be useful to teams working with GitHub Copilot Agent Mode and broader agentic development workflows.

The project is called Steering Studio:
https://github.com/ejochens/steering-studio

It is a standalone web application that helps users create a strong project starter pack before asking AI tools to generate code. The app guides users through structured intake forms and AI-assisted conversation, then generates a set of markdown steering documents that capture product context, technical constraints, architecture direction, workflows, testing expectations, and development guidance.

The main idea is simple: many teams struggle to get strong results from AI coding tools when they start with incomplete context, weak requirements, or inconsistent prompt quality. Steering Studio is meant to help solve that by making the “project context and guidance” step more structured, reusable, and easier to produce.

I designed it so the output can support:

• GitHub Copilot Agent Mode / Visual Studio workflows
• Amazon Kiro steering-driven workflows
• teams that want a repeatable design-first and context-first approach before code generation

A few things that may be relevant to Copilot users:

• it aims to reduce ad hoc prompting by generating a cleaner baseline of project guidance
• it is intended to improve consistency for requirements, architecture, testing, and coding expectations
• it can help teams start new projects or major enhancements with better context for AI-assisted development
• it is designed as a practical bridge between human-led discovery and agent-assisted implementation

I’d really appreciate feedback from Copilot users and devs on questions like:

• Is this the kind of setup artifact that would improve real-world Copilot Agent Mode outcomes?
• What project-context documents or repository instructions are most useful in practice?
• What would make this more effective for Visual Studio / Copilot workflows?
• Are there better ways to structure generated guidance so it is more usable by Copilot-powered development sessions?

The repo is public and I’d welcome suggestions, criticism, or ideas for improvement:
https://github.com/ejochens/steering-studio

[github-actions[bot]]

💬 Your Product Feedback Has Been Submitted 🎉

Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users.

Here's what you can expect moving forward ⏩

• Your input will be carefully reviewed and cataloged by members of our product teams. 
  • Due to the high volume of submissions, we may not always be able to provide individual responses.
  • Rest assured, your feedback will help chart our course for product improvements.
• Other users may engage with your post, sharing their own perspectives or experiences. 
• GitHub staff may reach out for further clarification or insight. 
  • We may 'Answer' your discussion if there is a current solution, workaround, or roadmap/changelog post related to the feedback.

Where to look to see what's shipping 👀

• Read the Changelog for real-time updates on the latest GitHub features, enhancements, and calls for feedback.
• Explore our Product Roadmap, which details upcoming major releases and initiatives.

What you can do in the meantime 💻

• Upvote and comment on other user feedback Discussions that resonate with you.
• Add more information at any point! Useful details include: use cases, relevant labels, desired outcomes, and any accompanying screenshots.

As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities.

Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐

[7Vedansh]

This is a really interesting idea it feels like you’re solving a problem most people don’t explicitly talk about bad or incomplete context going into AI tools rather than the output itself.

I like the “starter pack” approach. In practice, Copilot (especially in agent-style workflows) tends to perform much better when things like architecture direction, constraints, and testing expectations are clearly defined upfront otherwise it just fills gaps with assumptions.

Also, how the content is structured could matter a lot. Models usually respond better to clear, predictable sections and constraints rather than long-form explanations. Maybe experimenting with more structured or “agent-friendly” formats could improve results further.

On your core question, I think yes but only if the context is actually used during development and it’s concise enough for the model to reliably consume

[ejochens]

I agree that proper starting context is invaluable to success and reduced issues. I agree that more refinement in the output of the steering documents is necessary. Please try it out and see if it helps you, I would love to hear your results!

[7Vedansh]

I gave this a quick try on a small project with Copilot, and it actually helped with the initial output especially around structure and constraints. It felt like Copilot made fewer random assumptions. One thing I noticed though shorter, more structured sections worked better than long explanations. Also, as things evolved, I wanted a way to update parts of the context instead of regenerating everything.

Overall, really useful idea — curious how you’re thinking about handling updates over time?