feat: decision log guidelines and start

[KayTV]

Checklist

• Update PR Title to follow this pattern: [INTENT]:[MESSAGE]

  The title will become a one-line commit message in the git log, so be as concise and specific as possible -- refer to How to Write a Git Commit Message. Prepend Conventional Commit intent (fix:, feat:, chore:, ci:, docs:, style:, refactor:, perf:, test:).

Ticket

Resolves #{TICKET NUMBER or URL or description} or Adds {new capability or feature}

Changes

This pull request adds a new architectural decision record to the documentation, outlining the evaluation of frontend UI framework options for implementing a chat interface for the Mastra bot. The document compares a custom-built UI approach with a template-based solution, highlighting the pros and cons of each, and provides context, decision drivers, and links to relevant technical stories.

Documentation: Decision Record for Frontend UI Framework

• Added docs/decisions/ui-framework/000-custom-built-ui-vs-template-chat.md to document the evaluation of custom-built UI versus template-based approaches for the Mastra bot chat interface, including decision drivers, considered options, pros and cons, and links to related PRs and resources.

Context for reviewers

Decision log for tracking progress of application

[fg-nava]

For 3P integrations, I would recommend further looking into and surfacing these as options we reviewed, considered, or ruled-out:

• https://shadcn-chatbot-kit.vercel.app/
• https://www.langui.dev/
• any others you may have come across

I think we could then briefly expand on the "why" behind each good and bad in the Pros+Cons section for 3P integrations below.

[fg-nava]

This would also be a consequence of Option 1, right?

Bad, because requires integration with existing Mastra backend

As you built out the initial phase for each solution where did the integration requirements differ?

[fg-nava]

Was this an option you considered? Is the difference between this option and Option 1 that it does not use AI SDK?

[KayTV]

Oh, my edits didn't save one sec

[fg-nava]

This is a good start! I liked how Christine and Jillian asked questions around some of the strategic differences. Would like if we could tailor the pros and cons more towards these.

• Maintenance burden vs. development speed
• Customization flexibility vs. out-of-the-box functionality
• Short-term demo needs vs. long-term project architecture
• Quality of existing templates vs. building it right from scratch
• Hybrid approach (Christine's "Frankenstein" mention) vs. picking one path