Magic Kick

The UI of my AI-Business-OS and my daily planner: tasks, goals, projects, schedule, journal, Firebase sync, AI coaching, summaries, and workflow automation.

Magic Kick is not a commercial product. It is a private-use internal tool that I keep public as implementation proof: a realistic, working system for managing personal execution, project flow, scheduling, reflection, and AI-assisted routines.

It demonstrates the kind of product and engineering work behind SME internal tools:

• turning scattered work into structured modules
• modeling goals, projects, tasks, schedule blocks, resources, and journal entries
• building dashboard-style operational visibility
• adding AI assistance where it supports review, planning, coaching, and summaries
• maintaining a documented AI-assisted development workflow

---

Public Proof Role

Magic Kick supports my positioning as an AI Systems Consultant and Technical Product Manager.

It is best understood as:

Internal tool proof
+ AI-assisted workflow proof
+ delivery-system sandbox

It is not positioned as:

SaaS product
consumer productivity app
client-ready SME solution

The value is in the implementation evidence: data models, modules, auth, sync, AI routes, UI density, workflow rules, and iterative delivery discipline.

---

What This Proves

Internal tools and dashboards

Magic Kick contains a modular command center for goals, projects, schedule, tasks, resources, journal, achievements, and execution feedback.

This supports the same delivery pattern used in SME internal tools:

• identify the operational objects
• model their states and relationships
• make ownership, dates, progress, and risk visible
• reduce scattered tracking across notes, calendars, and spreadsheets

AI workflow implementation

The app includes AI-oriented surfaces and backend routes for:

• coaching
• schedule suggestions
• retrospective summaries
• weekly summaries
• insights and risk/pattern logic

AI is not treated as magic text generation. It is wrapped around real workflow context and bounded by existing data structures.

AI-SDLC practice

The repository is also a sandbox for practicing AI-assisted software delivery:

Research -> design discussion -> notes -> GitHub issue -> implementation -> PR -> review log

Recent work shows this pattern in practice through scoped project and schedule improvements, ADRs, changelog entries, and UI iteration.

---

Current Status

Magic Kick is the UI of AI-Business-OS and a personal daily planner (ADR-024). It follows a ranked roadmap in docs/CONTROL_PLANE_UI_SPEC.md.

The product surface is intentionally capped at the existing modules:

1. Command Center
2. Goals
3. Todo
4. Projects
5. Achievements
6. Schedule
7. Resources
8. Journal
9. XP / Levels

Future work is limited to:

• personal-use fixes
• scoped AI-workflow experiments
• refinement inside existing modules
• documentation of decisions and learning

No broad SaaS roadmap is planned.

---

Main Capabilities

Area	Capability
Command Center	Daily operating surface for priorities, progress, and execution feedback
Goals	Strategic direction and longer-term outcomes
Projects	Project identity, status, dates, links, milestones, detail panels, and workflow visibility
Todo	Backlog, daily focus, task states, due dates, categories, and schedule-linked tasks
Schedule	Time blocks, editable notes, linked tasks, category metadata, due-date scheduling and unscheduling
Journal	Reflection and retrospective capture
Resources	Personal knowledge and reusable references
Achievements / XP	Motivation and feedback loop
AI Layer	Coaching, planning, summaries, risk/pattern support, and workflow review experiments

---

Tech Stack

Layer	Choice
Framework	Next.js 16 App Router + React 19 + TypeScript
Styling	Tailwind CSS v4 + shadcn/ui + Radix UI
State	Zustand with localStorage persistence
Backend	Firebase Auth + Cloud Firestore
AI	Anthropic SDK through server routes
Analytics	Vercel Analytics
Tooling	npm, TypeScript, ESLint, GitHub Actions

---

Architecture Snapshot

Next.js App Router
  -> authenticated app shell
  -> module surfaces
  -> Zustand local store
  -> localStorage persistence
  -> Firestore sync
  -> AI API routes
  -> documented AI-SDLC workflow

The app is offline-first from the UI perspective. Firestore is the remote sync target, while local state keeps the app responsive for personal daily use.

---

Documentation

Document	Purpose
Architecture	Current system shape, module relationships, sync, and AI layer
Decisions Log	ADR-style record of important design and governance choices
Firebase Architecture	Auth, Firestore, sync, and data-boundary details
Workflow Automation Playbook	How AI-assisted workflow experiments are scoped and reviewed
Scope Rules	Governing rule, module ceiling, and what a change can be
Daily Checklist	Daily operating checklist
Changelog	User-facing and proof-of-work history

---

Setup

Prerequisites

• Node.js 20+
• npm
• Firebase project credentials for local development

Commands

npm install
npm run dev
npm run build
npm run typecheck
npm run lint
npm run test:unit

Environment

1. Copy .env.example to .env.local.
2. Fill in the Firebase public client values.
3. Add AI-related server environment values only when testing AI routes locally.
4. Optional for emulator usage:

NEXT_PUBLIC_FIREBASE_USE_EMULATOR=true

OS canon feed (P7)

Command Center shows current focus from the private AI-Business-OS repo under Daily Focus, marked "From canon" (ADR-031). A server route (GET /api/os/canon, signed-in users only) reads two tracked files on main through the GitHub API, cached for 5 minutes:

• 01_CONTEXT/current-focus.md: the primary outcomes (bold titles) and the execution-allocation lanes at primary and protected, plus its reconciled_with_git date.

• 00_HOME/context-pack.md: only the generated date from its front matter, shown as "Canon as of , reconciled ".

• Env var: AIOS_GITHUB_TOKEN, a fine-grained PAT for romahawk/AI-Business-OS only with Contents: read. Server-side only; it is never sent to the browser or logged.

• Read-only. Nothing from canon is written to Firestore, the local store or sync. The block has no actions and does not count against the Daily Focus limit.

• States: no token shows nothing; no current-focus.md shows "Current focus not found in canon"; GitHub down or refusing shows the last good result as "Last known, fetched ", or one muted line if nothing was fetched yet; an unreadable context pack only drops the date stamp.

Close day: write-back to the OS (P11)

"Close day" on the Command Center writes today's daily review to AI-Business-OS (ADR-032), the Magic Kick side of the OS daily-close workflow. The dialog has the sections of the OS 09_TEMPLATES/daily-review-template.md. Completed starts with today's completed tasks and Carry Forward with the open Daily Focus tasks. Decisions go under "Decisions Made". The whole file is shown before anything is written.

• Route: GET /api/os/daily-review (does today's file exist?) and POST (create it), signed in as the writer only.
• Writes one file, 06_REVIEWS/daily/YYYY-MM-DD.md (Europe/Berlin date), directly to OS main with the message daily-close: YYYY-MM-DD (Magic Kick). The server refuses every other path. An existing file is never overwritten.
• Env vars: AIOS_GITHUB_WRITE_TOKEN (a separate fine-grained PAT, romahawk/AI-Business-OS only, Contents: read and write) and AIOS_WRITER_UID (Roman's Firebase uid). Server-side only. Without either, the dialog says Close day is not set up.
• Nothing from the dialog is kept in Firestore or the local store.

Local Development

npm run emulators
npm run dev

Open http://localhost:3000.

---

Deployment

• Web app: Vercel
• Auth and data: Firebase Auth + Cloud Firestore
• Production URL: not published in this repository

The app is public as source code, not as a shared production service.

---

Workflow

• Create a focused branch from the active base.
• Keep commits scoped and use type(scope): short description.
• Run npm run lint, npm run typecheck, and npm run build before merging.
• Update CHANGELOG.md for user-facing changes.
• Record meaningful product/architecture decisions in docs/DECISIONS_LOG.md.

---

Suggested Repository Metadata

Recommended GitHub description:

The UI of my AI-Business-OS and my daily planner: tasks, goals, projects, schedule, journal, Firebase sync, AI coaching, summaries, and workflow automation.

Recommended topics:

internal-tools
ai-workflows
workflow-automation
productivity-system
nextjs
typescript
firebase
firestore
ai-assisted-development
technical-product-management
dashboard
proof-of-work

---

License

Private personal project. All rights reserved.