- initialize Next.js app with TypeScript and Tailwind, #1
- set up GitHub Kanban board with Backlog / In Progress / Done columns, #2
- install Monaco Editor and render in basic page with no configuration, #4
- add language selector with boilerplate snippets, #8
- confirm Monaco renders TypeScript file with syntax highlighting, #5
- validate editor is functional: typing, cursor movement, basic keybindings, #6
- deploy editor to Vercel and confirm live URL, #3
- build three-panel IDE layout with file tree, Monaco editor, and AI panel, #12
- add custom tailwind classes and implement three-theme toggle, #17
- implement static file tree with hardcoded file list, #13
- wire file tree clicks to load placeholder text into Monaco editor, #14
- build AI panel UI shell with input, send button, and message list, #15
- enhance styling and finalize layout at desktop screens sizes, #16
- add /learn route and dashboard ui rendering first teaching module, #19
- define module schema and author module content for simple spa app, #20
- wire selection to load module files into the editor and file tree, #21
- extract dashboard, project, page headers as standalone components, #23
- add App Router navigation between editor and dashboard interfaces, #22
- content expansion and UI enhancements for learning modules, #24
- install Vercel AI SDK, create new IAM service user, store env vars, #27
- create POST /api/chat route with streaming response, #28
- replace mock response in AssistantPanel with useChat hook, #29
- track cursor position in CodeEditor and lift to page state, #30
- track highlighted text in CodeEditor and lift to page state, #36
- inject active file content and cursor line into each chat request, #31
- track user edits in memory and synchronize AI context injection, #37
- track selection bounds and conditionally inject highlighted text, #38
- add filename and selection bounds indicator to AI panel, #32
- add persona selection and tune response, refine format guidelines, #33
- measure time to first token and reduce response latency, #35
- add persona display component rendering longer role description, #44
- add model select and redesign chat input component, #39 (refs #34)
- consolidate color palette and design system, #40 (refs #34)
- finalize v2 colors for all 3 themes and polish visual hiearchy, #41 (refs #34)
- extend modules data, restyle dashboard, and polish projects page, #45 (refs #34)
- design system, chat input, and dashboard polish for AI Eng Singapore 2026, #34
- reorg learning content file structure and import/export logic, #48
- refactor three-panel IDE for responsiveness at mobile sizes, #42
- generate v1 content for all skills and projects, #47
- redesign learn dashboard with refactored HyperUI components, #51
- add assets and apply Codalect rebranding across all surfaces, #62
- replace Persona type with Task and ConstraintLevel data layer, #56
- redesign system prompt with Socratic base layers and task overlays, #57
- add TaskSelect pills and refactor PersonaSelect into ConstraintSelect, #58
- document eval methodology and run manual per-mode baseline evals, #59
- (Closed as will not complete), #60
- publish manual eval results and benchmark against baseline models, #61
- install @codesandbox/sandpack-react and confirm no conflicts, #67
- migrate file state from page.tsx Record to Sandpack files store, #68
- refactor assistant file/cursor/selection context for Sandpack store, #69
- build slide-out preview + console drawer with branded toggle, #70
- refactor lesson modules, test Nodebox, and update load pipeline, #75
- add expandable folders UI to FileTree for component nesting, #76
- build a themed, runnable React starter for the default sandbox, #77
- wire Tailwind into Sandpack preview and fix console log capture, #72
- verify Sandpack preview renders cleanly on Vercel preview deploy, #73
- final enhancements to Codalect before merging Sprint 6 into main, #78
- wire up model select component to backend API routes, #49
- add inline create, rename, delete to file tree against Sandpack store, #71