This is a passion project I created to experiment with integrating AI code assistants into my development process. It brings together two of my favorite mediums as a maker (yarn and software) and I enjoy tinkering with it as time allows.
The app allows for four main types of content:
- Yarn
- Projects
- Plans
- Swatches
Plans allow you to collect sources of inspiration and ideas including images, links, yarn from your stash, or swatches from your library. You can also link to projects from a plan once you're ready to start creating a pattern.
Projects should really be called patterns, and one of my goals is to eventually refactor the code and database schema to make this change. For now, just know that projects = patterns. They have an "edit" and a "make" mode, so you can draft a pattern and then easily change to make mode to track progress on a completed pattern as you work. The system will save your progress.
- Database: Supabase
- Code: Next.js, TypeScript, Tailwind CSS
I've experimented with various AI code assistants throughout this process, but for now Cursor is my tool of choice. There's a Cursor rules folder in the main directory that includes Markdown files created by Ryan Carson that he graciously open sourced and can be accessed here. Watch his interview on the How I AI podcast to lean how to use them in your process.
You can reproduce the same checks that run in CI:
# Install dependencies
npm ci
# 1. Static analysis (ESLint & TypeScript strict)
npm run analysis:static
# 2. Dependency vulnerability scan
npm run dep-scan
# 3. Summarize findings
npm run summary:prss
# 4. Convert the summary into a task checklist
npx tsx scripts/convert-summary-to-tasks.tsAll generated reports live in the reports/ directory, and the remediation checklist is written to tasks/TASKS.md.