infra(04-00): route /c/ → /client/, Dockerfile, Gitea deploy

- Rename src/app/c/[token] → src/app/client/[token]
- Update proxy.ts, ClientRow, admin client detail with /client/ path
- Add output: "standalone" to next.config.ts for Docker build
- Add Dockerfile (multi-stage, node:20-alpine) and .dockerignore
- Push schema to Coolify Postgres via SSH tunnel (drizzle-kit push ✓)
- Update CLAUDE.md constraint 4 to reflect /client/ route
- Add Phase 4 planning artifacts (04-00, 04-RESEARCH, 04-PATTERNS)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-05-21 16:12:05 +02:00
parent 49ef45da83
commit 5bf5dfce71
21 changed files with 3164 additions and 63 deletions
+19 -4
View File
@@ -30,7 +30,17 @@
| CAT-01 | File/database dei servizi con prezzi e cosa è incluso | Pending |
| CAT-02 | Usato come base per la generazione assistita dei preventivi | Pending |
## Flusso Claude (v2 — deferred to Phase 4)
## Progetti Multi-Project (Phase 4)
| ID | Requirement | Status |
|----|-------------|--------|
| PROJ-01 | Ogni cliente può avere N progetti; ogni progetto ha workspace indipendente (fasi, pagamenti, preventivo, timer) | Pending |
| PROJ-02 | La dashboard cliente mostra tabs per 2+ progetti; con 1 progetto mostra direttamente il workspace senza selettore | Pending |
| PROJ-03 | La pagina /admin/projects elenca tutti i progetti con €/h reale e timer; /admin/projects/[id] è il workspace progetto | Pending |
| PROJ-04 | Il link cliente supporta slug personalizzabile (/c/mario-rossi) con fallback al token esistente | Pending |
| PROJ-05 | Analytics profittabilità per progetto: ore lavorate, accepted_total, €/h reale vs target_hourly_rate globale | Pending |
## Flusso Claude (v2 — deferred to Phase 5)
| ID | Requirement | Status |
|----|-------------|--------|
@@ -64,6 +74,11 @@
| CAT-01 | Phase 3 | Pending |
| CAT-02 | Phase 3 | Pending |
| ADMIN-03 | Phase 3 | Pending |
| CLAUDE-01 | Phase 4 (v2) | Deferred |
| CLAUDE-02 | Phase 4 (v2) | Deferred |
| CLAUDE-03 | Phase 4 (v2) | Deferred |
| PROJ-01 | Phase 4 | Pending |
| PROJ-02 | Phase 4 | Pending |
| PROJ-03 | Phase 4 | Pending |
| PROJ-04 | Phase 4 | Pending |
| PROJ-05 | Phase 4 | Pending |
| CLAUDE-01 | Phase 5 (v2) | Deferred |
| CLAUDE-02 | Phase 5 (v2) | Deferred |
| CLAUDE-03 | Phase 5 (v2) | Deferred |
+2 -1
View File
@@ -89,8 +89,9 @@ Decimal phases appear between their surrounding integers in numeric order.
3. La pagina /admin/projects elenca tutti i progetti con €/h calcolato e bottone timer play/stop
4. Il link cliente supporta slug personalizzato (/c/mario-rossi) con fallback al token; slug impostabile da /admin/clients/[id]/edit
5. Il tab Timer di ogni progetto mostra analytics profittabilità: ore lavorate, accepted_total, €/h reale vs target_hourly_rate globale
**Plans**: 4 plans
**Plans**: 5 plans
**Plan list**:
- [ ] 04-00-PLAN.md — Infra: Postgres su Coolify + /c/ → /client/ rename + Dockerfile + hub.iamcavalli.net [RUN FIRST]
- [ ] 04-01-PLAN.md — Schema migration (projects, slug, settings, FK migration) + drizzle-kit push + query layer
- [ ] 04-02-PLAN.md — Admin projects list (/admin/projects) + ProjectRow + client detail project cards
- [ ] 04-03-PLAN.md — Admin project workspace (/admin/projects/[id]) + TimerTab + ProfitabilityCard + /admin/impostazioni
+7 -7
View File
@@ -4,14 +4,14 @@ milestone: v1.0
milestone_name: milestone
status: executing
stopped_at: Phase 1 execution complete — all 5 plans done, E2E verified (valid token 200, invalid 404)
last_updated: "2026-05-19T21:12:54.673Z"
last_activity: 2026-05-19
last_updated: "2026-05-21T11:56:14.461Z"
last_activity: 2026-05-21 -- Phase 4 planning complete
progress:
total_phases: 4
total_phases: 5
completed_phases: 3
total_plans: 13
total_plans: 17
completed_plans: 13
percent: 100
percent: 76
---
# Project State
@@ -27,8 +27,8 @@ See: .planning/PROJECT.md (updated 2026-05-09)
Phase: 4
Plan: Not started
Status: Executing Phase 03
Last activity: 2026-05-19
Status: Ready to execute
Last activity: 2026-05-21 -- Phase 4 planning complete
Progress: [██░░░░░░░░] 25%
+15
View File
@@ -0,0 +1,15 @@
{
"mode": "yolo",
"granularity": "coarse",
"parallelization": true,
"commit_docs": false,
"model_profile": "budget",
"workflow": {
"research": true,
"plan_check": true,
"verifier": true,
"nyquist_validation": false,
"auto_advance": false,
"_auto_chain_active": false
}
}
@@ -0,0 +1,371 @@
---
phase: 04-progetti-multi-project
plan: "00"
type: execute
wave: 0
depends_on: []
files_modified:
- src/app/client/[token]/layout.tsx
- src/app/client/[token]/page.tsx
- src/proxy.ts
- src/components/admin/ClientRow.tsx
- src/app/admin/clients/[id]/page.tsx
- Dockerfile
- .env.example
autonomous: false
must_haves:
truths:
- "La route /client/[token] esiste e risponde correttamente (cartella rinominata da /c/)"
- "src/proxy.ts controlla /client/ invece di /c/ nel guard e nel matcher"
- "ClientRow.tsx e admin/clients/[id]/page.tsx usano /client/ nei link generati"
- "Il Dockerfile produce un build Next.js funzionante (exit code 0)"
- "DATABASE_URL in .env punta al Postgres self-hosted su Coolify (non Neon)"
- "drizzle-kit push sul nuovo DB completes con exit code 0"
- "L'app risponde su hub.iamcavalli.net dopo il deploy su Coolify"
artifacts:
- path: "src/app/client/[token]/page.tsx"
provides: "Dashboard cliente alla nuova route"
contains: "export default"
- path: "src/proxy.ts"
provides: "Middleware aggiornato per /client/"
contains: "pathname.startsWith(\"/client/\")"
- path: "Dockerfile"
provides: "Container image per Coolify"
contains: "FROM node:"
key_links:
- from: "src/proxy.ts"
to: "src/app/client/[token]/"
via: "matcher: ['/admin/:path*', '/client/:path*']"
pattern: "/client/:path*"
- from: "src/components/admin/ClientRow.tsx"
to: "src/app/client/[token]/"
via: "href={`/client/${client.token}`}"
pattern: "/client/"
---
<objective>
Infrastruttura pre-Phase 4: migrazione da Neon a Postgres self-hosted su Coolify (Hetzner), rinomina route cliente da /c/ a /client/, Dockerfile per deploy, dominio hub.iamcavalli.net.
Questo piano esegue PRIMA di 04-01 per garantire che tutto il codice Phase 4 venga scritto e testato sull'infrastruttura finale. Nessun refactor post-esecuzione.
Output: App deployata su hub.iamcavalli.net con DB Postgres su Coolify, route /client/[token] funzionante.
</objective>
<context>
@/Users/simonecavalli/IAMCAVALLI/.planning/ROADMAP.md
@/Users/simonecavalli/IAMCAVALLI/CLAUDE.md
<interfaces>
<!-- Riferimenti correnti da aggiornare -->
proxy.ts riga 32: `if (pathname.startsWith("/c/"))`
proxy.ts riga 33: `pathname.match(/^\/c\/([a-zA-Z0-9_-]+)/)`
proxy.ts riga 63: `matcher: ["/admin/:path*", "/c/:path*"]`
ClientRow.tsx riga 59: `href={\`/c/${client.token}\`}`
admin/clients/[id]/page.tsx riga 46: `href={\`/c/${client.token}\`}`
Cartella da rinominare: src/app/c/ → src/app/client/
</interfaces>
</context>
<threat_model>
T-00-01: Redirect loop dopo rinomina — mitigazione: aggiornare proxy.ts prima di rinominare la cartella, verificare matcher.
T-00-02: DB connection string esposta — mitigazione: .env mai committato, .env.example con placeholder.
T-00-03: Dati di test persi durante migrazione DB — accettabile: tutti i dati attuali sono test data (CONTEXT.md).
T-00-04: Coolify deploy fallisce silenziosamente — mitigazione: verificare health check dopo deploy.
</threat_model>
<tasks>
<task id="1" type="execute" autonomous="false">
<name>Task 1: Postgres su Coolify + migrazione DATABASE_URL</name>
<read_first>
- .env (connection string Neon attuale — non committare)
- drizzle.config.ts (per confermare che legge DATABASE_URL)
</read_first>
<action>
**A. Crea Postgres su Coolify (manuale — non automatizzabile)**
Nel pannello Coolify su Hetzner:
1. New Resource → Database → PostgreSQL
2. Nome: `clienthub-db`
3. Version: 16
4. Salva le credenziali generate (host, port, user, password, dbname)
5. Abilita "Public port" se necessario per drizzle-kit push da locale
**B. Aggiorna .env locale**
Sostituisci la riga DATABASE_URL con la connection string Coolify:
```
DATABASE_URL=postgresql://USER:PASSWORD@HOST:PORT/DBNAME?sslmode=require
```
Mantieni la vecchia riga commentata come backup:
```
# DATABASE_URL=postgresql://neon... (backup)
```
**C. Aggiorna .env.example**
Sostituisci il placeholder Neon con quello generico:
```
DATABASE_URL=postgresql://user:password@host:5432/dbname?sslmode=require
```
**D. Push schema al nuovo DB**
```bash
npx drizzle-kit push
```
Il DB è fresh — nessuna migrazione soft necessaria, tutti i dati attuali sono test data.
</action>
<verify>
<automated>npx drizzle-kit push 2>&1; echo "Exit: $?"</automated>
</verify>
<acceptance_criteria>
- drizzle-kit push completes con exit code 0
- .env contiene DATABASE_URL che punta a Coolify (non neon.tech)
- .env.example non contiene credenziali reali
- npx tsc --noEmit non produce errori legati al DB
</acceptance_criteria>
</task>
<task id="2" type="execute" autonomous="true">
<name>Task 2: Rinomina route /c/ → /client/ + aggiorna reference</name>
<read_first>
- src/proxy.ts (middleware completo — LEGGERE PRIMA di modificare)
- src/components/admin/ClientRow.tsx (link generato riga 59)
- src/app/admin/clients/[id]/page.tsx (link generato riga 46)
- src/app/c/[token]/page.tsx (dashboard cliente da spostare)
- src/app/c/[token]/layout.tsx (layout da spostare)
</read_first>
<action>
**A. Rinomina cartella App Router**
```bash
mv src/app/c src/app/client
```
Questo sposta automaticamente page.tsx e layout.tsx alla nuova route.
**B. Aggiorna src/proxy.ts**
Tre sostituzioni esatte:
Riga 32 — cambia:
```typescript
if (pathname.startsWith("/c/")) {
```
in:
```typescript
if (pathname.startsWith("/client/")) {
```
Riga 33 — cambia:
```typescript
const tokenMatch = pathname.match(/^\/c\/([a-zA-Z0-9_-]+)/);
```
in:
```typescript
const tokenMatch = pathname.match(/^\/client\/([a-zA-Z0-9_-]+)/);
```
Riga 63 — cambia:
```typescript
matcher: ["/admin/:path*", "/c/:path*"],
```
in:
```typescript
matcher: ["/admin/:path*", "/client/:path*"],
```
**C. Aggiorna ClientRow.tsx**
Riga 59 — cambia:
```typescript
href={`/c/${client.token}`}
```
in:
```typescript
href={`/client/${client.token}`}
```
**D. Aggiorna admin/clients/[id]/page.tsx**
Riga 46 — cambia:
```typescript
href={`/c/${client.token}`}
```
in:
```typescript
href={`/client/${client.token}`}
```
</action>
<verify>
<automated>grep -r '"/c/' src/ --include="*.ts" --include="*.tsx" 2>&1; echo "---"; grep -r "'/c/" src/ --include="*.ts" --include="*.tsx" 2>&1; echo "---"; grep -r '/c/:path' src/ --include="*.ts" --include="*.tsx" 2>&1</automated>
</verify>
<acceptance_criteria>
- `src/app/client/[token]/page.tsx` esiste (ls conferma)
- `src/app/c/` NON esiste più (ls conferma)
- `grep '"/c/' src/ -r` produce zero risultati
- `grep "startsWith(\"/client/\")" src/proxy.ts` produce un match
- `grep "/client/:path\*" src/proxy.ts` produce un match
- `npx tsc --noEmit` exit code 0
</acceptance_criteria>
</task>
<task id="3" type="execute" autonomous="true">
<name>Task 3: Dockerfile per Coolify</name>
<read_first>
- package.json (script build, versione Node richiesta)
- next.config.ts (per verificare se output: standalone è già presente)
- .gitignore (per confermare che .env non è committato)
</read_first>
<action>
**A. Aggiungi output standalone a next.config.ts**
Next.js standalone mode produce un bundle minimale ottimale per Docker.
In next.config.ts, aggiungi dentro l'oggetto config:
```typescript
output: "standalone",
```
**B. Crea Dockerfile nella root del progetto**
```dockerfile
FROM node:20-alpine AS base
FROM base AS deps
RUN apk add --no-cache libc6-compat
WORKDIR /app
COPY package.json package-lock.json* ./
RUN npm ci
FROM base AS builder
WORKDIR /app
COPY --from=deps /app/node_modules ./node_modules
COPY . .
RUN npm run build
FROM base AS runner
WORKDIR /app
ENV NODE_ENV=production
ENV NEXTAUTH_URL=https://hub.iamcavalli.net
RUN addgroup --system --gid 1001 nodejs
RUN adduser --system --uid 1001 nextjs
COPY --from=builder /app/public ./public
COPY --from=builder --chown=nextjs:nodejs /app/.next/standalone ./
COPY --from=builder --chown=nextjs:nodejs /app/.next/static ./.next/static
USER nextjs
EXPOSE 3000
ENV PORT=3000
ENV HOSTNAME="0.0.0.0"
CMD ["node", "server.js"]
```
**C. Crea .dockerignore nella root**
```
.env
.env.local
node_modules
.next
.git
.planning
```
**D. Aggiungi variabili d'ambiente su Coolify**
Nel pannello Coolify → app → Environment Variables, aggiungere:
```
DATABASE_URL=postgresql://... (la stessa del .env locale)
NEXTAUTH_SECRET=...
NEXTAUTH_URL=https://hub.iamcavalli.net
ADMIN_EMAIL=...
ADMIN_PASSWORD=...
```
**E. Configura dominio su Coolify**
Nel pannello Coolify → app → Domains:
- Aggiungi: `hub.iamcavalli.net`
- Abilita SSL automatico (Let's Encrypt)
</action>
<verify>
<automated>docker build -t clienthub-test . 2>&1 | tail -5; echo "Exit: $?"</automated>
</verify>
<acceptance_criteria>
- `Dockerfile` esiste nella root del progetto
- `.dockerignore` esiste e contiene `.env`
- `next.config.ts` contiene `output: "standalone"`
- `docker build` completes con exit code 0 (se Docker disponibile localmente)
- In alternativa: `npm run build` exit code 0 (verifica il build Next.js)
- Il file `.env` NON appare in `git status` (confermato da .gitignore)
</acceptance_criteria>
</task>
<task id="4" type="checkpoint" subtype="human-verify" autonomous="false">
<name>Checkpoint: verifica deploy su hub.iamcavalli.net</name>
<what_was_built>
- Postgres self-hosted su Coolify con schema applicato
- Route /client/[token] funzionante (rinominata da /c/)
- Dockerfile per deploy su Coolify
- Dominio hub.iamcavalli.net configurato
</what_was_built>
<how_to_verify>
1. Apri https://hub.iamcavalli.net — deve rispondere (anche con pagina 404 standard Next.js)
2. Apri https://hub.iamcavalli.net/admin/login — deve mostrare la schermata di login
3. Fai login come admin — deve accedere alla lista clienti
4. Copia il link di un cliente — deve essere nella forma `hub.iamcavalli.net/client/[token]`
5. Apri il link cliente — la dashboard deve caricare correttamente
6. Verifica che il vecchio link `/c/[token]` restituisca 404 (non più attivo)
</how_to_verify>
<resume_signal>
Digita "hub ok" quando tutti i controlli passano.
</resume_signal>
</task>
</tasks>
<verification>
1. `src/app/c/` non esiste — `src/app/client/[token]/` esiste
2. `grep -r '"/c/' src/` produce zero risultati
3. `grep "startsWith(\"/client/\")" src/proxy.ts` produce un match
4. `npx drizzle-kit push` exit code 0 sul nuovo DB
5. `npm run build` exit code 0
6. https://hub.iamcavalli.net risponde dopo deploy Coolify
7. https://hub.iamcavalli.net/client/[token] carica la dashboard cliente
</verification>
<summary_template>
## Plan 04-00 Complete
**Infrastruttura migrata:**
- DB: Neon → Postgres self-hosted su Coolify (Hetzner)
- Route: /c/[token] → /client/[token]
- Deploy: Vercel → Coolify via Docker
- Dominio: hub.iamcavalli.net attivo
**File modificati:** src/proxy.ts, src/components/admin/ClientRow.tsx, src/app/admin/clients/[id]/page.tsx, next.config.ts
**File creati:** Dockerfile, .dockerignore
**File spostati:** src/app/c/ → src/app/client/
</summary_template>
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,921 @@
# Phase 04: Progetti — Multi-Project per Cliente - Research
**Researched:** 2026-05-21
**Domain:** Data model refactoring + multi-tier architecture (DB schema migration, API routing, client/admin UI)
**Confidence:** HIGH
## Summary
Phase 04 transforms ClientHub from a single-project-per-client model to a multi-project model. This is a **breaking schema migration** where the `projects` table becomes the primary work container, and 6 existing tables (`phases`, `payments`, `quote_items`, `time_entries`, `documents`, `notes`) move their FK from `client_id` to `project_id`. The `clients` table gains a `slug` field and loses denormalized fields that move to the project level.
All existing data is test data — hard migration (drop/recreate tables) is acceptable and planned.
**Key insight:** This is NOT a typical multi-tenancy refactor. It's a structural deepening: clients now own projects, and projects own the work. The middleware routing pattern (`/c/[slug-or-token]`) stays the same, but resolves at the client level and then queries to find projects.
**Primary recommendation:** Use vertical slice approach (Wave 0 schema, Wave 1 core routing/queries, Wave 2 admin UI, Wave 3 client UI + analytics). All 5 locked architectural decisions are already finalized in CONTEXT.md — implement them as-is, no discretion needed.
---
## User Constraints (from CONTEXT.md)
### Locked Decisions
**Schema & Data Model**
- D-01: New `projects` table with `id`, `client_id` FK, `name` (brand/project name), `archived`, `created_at`. No direct `accepted_total` — denormalized from `quote_items` per project.
- D-02: Six tables move FK from `client_id``project_id`: `phases`, `payments`, `quote_items`, `time_entries`, `documents`, `notes`. `comments` stays polymorphic on `entity_id` (unchanged).
- D-03: `clients` table loses project-scoped fields. Retains: `id`, `name`, `brand_name`, `token`, `slug` (new), `archived`, `created_at`. `accepted_total` moves to `projects`.
- D-04: `slug` field added to `clients` — optional, unique, URL-safe (e.g., `mario-rossi`). Middleware tries slug first, falls back to token.
- D-05: `projects.accepted_total` denormalized (text, nullable), admin sets manually in project Preventivo tab.
**Link & Access**
- D-06: Token stays on `clients` for auth middleware. Middleware checks slug first (DB lookup), then token (existing pattern). Both grant access.
- D-07: Slug is set in `/admin/clients/[id]/edit` (new form field, optional, with link preview).
- D-08: Route `/c/[token-or-slug]` unchanged in path — middleware resolves both.
**Client Dashboard**
- D-09: 1 project → direct view (no selector).
- D-10: 2+ projects → tabs with brand names (shadcn Tabs, already in codebase).
- D-11: Project view identical to current client dashboard but scoped to one project.
**Admin — Client List View**
- D-12: `/admin/clients` shows client name + project brands as secondary text (e.g., "Mario Rossi" / "Brand Blu | Brand Verde"). LTV = sum of all project `accepted_total`.
- D-13: Clicking a client opens `/admin/clients/[id]` showing project cards/rows (not workspace directly).
**Admin — Project List & Workspace**
- D-14: New `/admin/projects` (NavBar link) — all projects with: Name, Parent Client, Value (accepted_total), Acconto, Saldo, Timer, €/h.
- D-15: Timer in projects list shows play/stop for each project. Only one timer active at a time (scoped to project now).
- D-16: `/admin/projects/[id]` workspace identical to current `/admin/clients/[id]` but project-level: Panoramica, Fasi, Documenti, Pagamenti, Note, Preventivo, Timer, Commenti tabs.
**Project Creation**
- D-17: Create project from: (1) `/admin/clients/[id]` with "+ Nuovo Progetto" button, or (2) `/admin/projects` with "+ Nuovo Progetto" → select client.
- D-18: Creation form: Project Name (brand) + Client (if from list). No other fields at creation time.
**Timer & Analytics**
- D-19: `time_entries.client_id``time_entries.project_id`. Timer now per-project.
- D-20: Analytics profittabilità in project Timer tab: total hours, accepted_total, €/h real (accepted ÷ hours), target rate × hours (ideal cost), delta (gain/loss vs target).
- D-21: `target_hourly_rate` is global (e.g., 50€/h) stored in `settings` table or env var. New "Impostazioni" page in NavBar for admin to set.
- D-22: Statistiche page shows aggregated profitability for all projects + breakdown per client.
### Claude's Discretion
- **Settings table structure:** key-value (simplest) vs. dedicated columns. Recommendation: `settings(key text PK, value text)`.
- **Tab order in project detail:** Follow current client detail order.
- **Project card style in client detail:** Reuse existing UI patterns.
### Deferred Ideas (OUT OF SCOPE)
- Automatic invoice generation per project.
- PDF export for quotes.
- AI Onboarding (Phase 5 — requires this phase as prerequisite).
- Email notifications when phases change.
---
## Phase Requirements
| ID | Description | Research Support |
|----|-------------|------------------|
| PROJ-01 | Every client can have N independent projects; each project has its own workspace (phases, payments, quote, timer) accessible from /admin/projects/[id] | Schema migration (D-01, D-02), Query refactor (getProjectFullDetail), Admin workspace pages (/admin/projects/[id]) |
| PROJ-02 | Client dashboard shows tabs for 2+ projects; 1 project shows direct workspace without selector | Client view refactor for multi-project detection, Tabs UI pattern (shadcn already available), Client page routing logic |
| PROJ-03 | /admin/projects lists all projects with €/h calculated and timer play/stop; /admin/projects/[id] is the project workspace | New project list page and detail page templates (clone from ClientRow/Client detail), Timer refactor (client_id → project_id) |
| PROJ-04 | Client link supports custom slug (/c/mario-rossi) with fallback to token; slug settable from /admin/clients/[id]/edit | Middleware slug resolution (internal API route for DB lookup), Clients edit form, Link preview component |
| PROJ-05 | Profitability analytics per project: hours tracked, accepted_total, €/h real vs target_hourly_rate global | Analytics card in Timer tab (formula: accepted ÷ hours = €/h real, target × hours = ideal cost, delta = profit/loss), Settings table for global target rate |
---
## Architectural Responsibility Map
| Capability | Primary Tier | Secondary Tier | Rationale |
|------------|-------------|----------------|-----------|
| Project CRUD (create, read, update, archive) | API / Backend | Admin Browser | Server actions + DB operations live in API tier; admin calls via form actions |
| Timer start/stop for projects | API / Backend | Admin Browser | Timer logic (duration calculation, active session tracking) is backend; UI shows state via server actions |
| Multi-project dashboard routing | Frontend Server (SSR) | Browser | Server chooses 1-project direct view vs. 2+ project tabs; browser renders tabs (Tabs component is client-side) |
| Slug lookup & resolution | API / Backend + Edge Middleware | — | Middleware calls internal API route to resolve slug → client_id; API accesses DB (can't do direct queries in Edge runtime) |
| Profitability analytics calculation | API / Backend | Admin Browser | Formula applied server-side (accepted_total ÷ duration_seconds), displayed in admin workspace |
| Client project visibility | API / Backend | Browser | Client API (`/c/[token]/*`) queries projects belonging to the resolved client; client browser renders what API returns |
---
## Standard Stack
### Core
| Library | Version | Purpose | Why Standard |
|---------|---------|---------|--------------|
| Next.js App Router | 16 | Meta-framework for API routes + SSR + auth middleware | Established in project; Edge middleware pattern for token/slug resolution |
| Neon Postgres | current | Primary database | Established; supports both pooled (API routes) and direct (drizzle-kit) connections |
| Drizzle ORM | current | Type-safe query builder + schema management | Established; `drizzle-kit push` handles migrations without manual SQL |
| Auth.js v4 | current | Admin session authentication | Established; `/admin/*` routes use Auth.js session guard |
| Tailwind v4 | current | Styling | Established; Tailwind scanning configured to include project source |
| shadcn/ui | current | UI components library | Established; Tabs component already used for admin workspaces |
| Zod | current | Input validation | Established for form validation |
| nanoid | current | Random ID generation | Established; used for all entity IDs |
### Supporting
| Library | Version | Purpose | When to Use |
|---------|---------|---------|-------------|
| react-hook-form | (check package.json) | Form state management | Forms in admin (create project, edit client slug) |
| @radix-ui/tabs | (check package.json) | Underlying Tabs component | shadcn Tabs wrapper — already available |
### Alternatives Considered
| Instead of | Could Use | Tradeoff |
|------------|-----------|----------|
| `settings` key-value table | Hardcode target_hourly_rate as env var | Env var: simpler, no DB call. Table: more flexible, admin can change via UI. Recommendation: start with table for flexibility. |
| Slug as optional field in clients | Always-require slug, generate from name | Optional is better: gradual migration, existing clients keep token links, new clients can have slug. |
| Clone workspace from client detail | Build project detail from scratch | Cloning is faster: tabs, layout, queries already proven. Reduces bugs. |
**Installation:**
```bash
# No new packages needed — all standard stack already installed
npm ls next neon drizzle-orm next-auth tailwindcss shadcn-ui zod nanoid
```
**Version verification:**
All versions are in the existing `package.json` and `drizzle.config.ts`. No new dependencies required for Phase 04 schema/routing. Any new UI components (if needed) are installed via `npx shadcn-ui@latest add [component]`.
---
## Architecture Patterns
### System Architecture Diagram
```
┌─────────────────────────────────────────────────────────────────────┐
│ CLIENT BROWSER │
│ ┌──────────────┐ ┌──────────────────────────────────────────────┐ │
│ │ /c/[slug] │ │ Client Dashboard (multi-project) │ │
│ │ or /c/[tok] │──│ ├─ 1 project: direct view │ │
│ │ │ │ └─ 2+: tabs per brand name │ │
│ └──────────────┘ │ ├─ Phases, Tasks, Deliverables │ │
│ │ ├─ Payments & Status │ │
│ │ └─ Documents & Notes │ │
│ └──────────────────────────────────────────────┘ │
│ ┌──────────────┐ ┌──────────────────────────────────────────────┐ │
│ │ /admin/* │ │ Admin Area (multi-workspace) │ │
│ │ (Auth.js) │──│ ├─ /admin/clients: list + LTV │ │
│ │ │ │ ├─ /admin/clients/[id]: projects cards │ │
│ └──────────────┘ │ ├─ /admin/projects: all projects + timer │ │
│ │ ├─ /admin/projects/[id]: workspace (tabs) │ │
│ │ └─ /admin/impostazioni: target_hourly_rate │ │
│ └──────────────────────────────────────────────┘ │
└──────────────────────────────────────────────────────────────────────┘
┌──────────────────────────┐
│ Edge Middleware │
│ (proxy.ts / middleware) │
│ ├─ Resolve /c/[slug] │
│ │ → call /api/internal/ │
│ │ validate-slug │
│ ├─ Fallback /c/[token] │
│ │ → existing pattern │
│ └─ Admin session check │
└──────────────────────────┘
┌──────────────────────────────────────────┐
│ Next.js API Routes (Node.js runtime) │
│ ├─ /api/internal/validate-token │
│ ├─ /api/internal/validate-slug (NEW) │
│ ├─ /api/auth/* (NextAuth) │
│ └─ Server Actions (form submissions) │
└──────────────────────────────────────────┘
┌──────────────────────────────────────────┐
│ Drizzle ORM Query Layer │
│ ├─ getAllProjectsWithPayments (NEW) │
│ ├─ getProjectFullDetail (NEW) │
│ ├─ getClientWithProjects (NEW) │
│ ├─ slugToClientId (NEW, for resolution) │
│ └─ timer-actions refactored │
│ (client_id → project_id) │
└──────────────────────────────────────────┘
┌──────────────────────────────────────────┐
│ Neon Postgres Database │
│ ├─ clients (+ slug field NEW) │
│ ├─ projects (NEW, client_id FK) │
│ ├─ phases (FK: project_id instead) │
│ ├─ tasks (unchanged, phase_id FK) │
│ ├─ payments (FK: project_id instead) │
│ ├─ quote_items (FK: project_id instead) │
│ ├─ time_entries (FK: project_id NEW) │
│ ├─ documents (FK: project_id instead) │
│ ├─ notes (FK: project_id instead) │
│ ├─ settings (NEW, key-value table) │
│ ├─ comments (entity_id, unchanged) │
│ ├─ deliverables (unchanged) │
│ └─ service_catalog (unchanged) │
└──────────────────────────────────────────┘
```
**Data flow for client dashboard:**
1. Browser requests `/c/mario-rossi`
2. Middleware intercepts, calls `/api/internal/validate-slug?slug=mario-rossi`
3. API resolves slug → client_id via `slugToClientId()`
4. Server component calls `getClientWithProjects(client_id)`
5. If 1 project: render direct project view (call `getProjectFullDetail(project_id)`)
6. If 2+: render tabs, each tab calls `getProjectFullDetail(project_id)`
7. Browser displays project workspace with phases, payments, documents, notes
**Data flow for admin projects list:**
1. Admin visits `/admin/projects`
2. Server calls `getAllProjectsWithPayments()`
3. Returns projects with: parent client name, accepted_total, payment statuses, active timer info, calculated €/h
4. Renders ProjectRow for each (clone of ClientRow pattern)
---
## Runtime State Inventory
> This phase involves renaming + moving FK relationships from `client_id` to `project_id`. Verify all runtime state.
| Category | Items Found | Action Required |
|----------|-------------|------------------|
| **Stored data** | Current DB has ~13 tables. Hard migration acceptable (drop/recreate from schema). Test data only — no customer data to preserve. | Code edit: `src/db/schema.ts` (new `projects` table, update FK on 6 tables, add `slug` to clients, new `settings` table). Execute `drizzle-kit push` to apply to Neon. |
| **Live service config** | No external service configuration (n8n workflows, webhooks, etc.) references client_id or project structure explicitly. | None — if services are added in future phases, ensure they use project_id. |
| **OS-registered state** | None — this is a web application with no local task scheduling or registered executables. | None required. |
| **Secrets/env vars** | `.env` currently has: DATABASE_URL, NEXTAUTH_SECRET, NEXTAUTH_URL. No references to client/project IDs. New `target_hourly_rate` will be stored in DB `settings` table (not env var). | None for secrets. If admins prefer env var, add `TARGET_HOURLY_RATE=50` to `.env` and read in analytics component. Recommendation: use DB table for flexibility. |
| **Build artifacts** | No build artifacts reference client_id or project structure. Next.js builds are stateless. | None required. Fresh build after schema migration. |
**Conclusion:** All data is test data. Hard migration is acceptable. No runtime state inventory concerns blocking execution.
---
## Common Pitfalls
### Pitfall 1: Incomplete FK Migration Across Tables
**What goes wrong:** Missing one table's FK migration (e.g., forgetting to update `time_entries.client_id``project_id`), causing orphaned records or failed queries in admin workspace when drilling into a specific project.
**Why it happens:** 6 tables need updating. Easy to miss one if checklist is informal.
**How to avoid:**
1. List all 6 tables in schema migration PR title: `phases, payments, quote_items, time_entries, documents, notes`.
2. After `drizzle-kit push`, verify each table with `\dt public.*` in psql and spot-check that old client_id FK is gone, new project_id FK is present.
3. In planner, create a Wave 0 schema-only task + verification subtasks for each table.
**Warning signs:**
- `getProjectFullDetail()` returns empty phases/payments even though they exist in the DB.
- Timer actions fail with "project_id not found" FK violation on insert.
### Pitfall 2: Client Middleware Resolution Order (Slug vs. Token)
**What goes wrong:** Middleware checks token first and finds a match before trying slug, so `/c/mario-rossi` is treated as an invalid token and returns 404 even though the slug exists.
**Why it happens:** Easy to reverse the order in `validate-slug` or middleware logic.
**How to avoid:**
1. Middleware must call `/api/internal/validate-slug?slug=...` FIRST.
2. Only if slug lookup fails (404 from API), fall back to existing token validation.
3. Document the order in code comment.
**Warning signs:**
- Slug links return 404 even though slug is in the DB and client can access via token link.
- Creating a new slug for an existing client breaks the old token link (should not).
### Pitfall 3: Admin Workspace Queries Not Scoped to Current Project
**What goes wrong:** `getProjectFullDetail()` accidentally returns data from multiple projects or from the wrong project due to missing WHERE clause on project_id.
**Why it happens:** Copy-pasting from `getClientFullDetail()` and forgetting to update the WHERE conditions.
**How to avoid:**
1. After writing `getProjectFullDetail()`, trace through each query: phases, tasks, deliverables, payments, documents, notes, comments, quote_items.
2. Verify each has `.where(eq(table.project_id, projectId))` or is a child query that's already filtered.
3. Add a comment above each query stating what it filters on.
**Warning signs:**
- Workspace shows phases/tasks from sibling projects.
- Clicking into a project workspace, then switching projects, shows the same data.
### Pitfall 4: Timer Still Checks for Global "Only One Active" Instead of Per-Project
**What goes wrong:** Admin starts timer for Project A, then clicks timer for Project B, and Project A's timer is stopped. User expects independent timers.
**Why it happens:** Current `startTimer()` stops ALL running sessions. Must be updated to allow one timer per project (or clarify that "global only one timer" is the design).
**How to avoid:**
1. Decision: Are timers per-project OR global (only one active per admin account)?
2. From CONTEXT (D-15): "Only one timer active at a time (scoped to project now)" suggests global is intended (one active total).
3. Keep current logic but verify: when admin starts project B's timer, project A's should auto-stop.
4. Add test: start timer A, start timer B, verify A is stopped and B is running.
**Warning signs:**
- Two projects have active timers simultaneously (duration_seconds null on both).
### Pitfall 5: Client Slug Field Validation Too Strict or Too Loose
**What goes wrong:** Slug regex rejects valid inputs (e.g., "mario-rossi-2") or accepts invalid ones (e.g., spaces, special chars).
**Why it happens:** Regex written without testing against edge cases.
**How to avoid:**
1. Define slug rule: lowercase alphanumeric + hyphens only, 3-50 chars, must be unique.
2. Zod schema: `slug: z.string().regex(/^[a-z0-9-]{3,50}$/).optional().or(z.null())`
3. Test form submission with: "mario-rossi", "mario-rossi-2", "MARIO" (should fail), "m--r" (ok?), "m" (too short), "m " (space, should fail).
**Warning signs:**
- Admin can't set a slug they expect to work (form rejects it).
- Middleware crashes on malformed slug from DB.
### Pitfall 6: "Settings" Table Key Mismatches in Code
**What goes wrong:** Code reads `settings.value WHERE key = 'hourly_rate'` but admin wrote it as `'target_hourly_rate'`, returning null and falling back to a hardcoded default.
**Why it happens:** Settings keys are strings with no schema enforcement. Easy to have typos or inconsistent naming.
**How to avoid:**
1. Define an enum or constant for all settings keys:
```typescript
const SETTINGS_KEYS = {
TARGET_HOURLY_RATE: 'target_hourly_rate',
} as const;
```
2. Always read via constant: `getSetting(SETTINGS_KEYS.TARGET_HOURLY_RATE)`.
3. Admin form submits value for this constant key only.
**Warning signs:**
- Analytics always shows a hardcoded rate (default value) instead of what admin set.
- Changing the setting has no effect.
---
## Code Examples
Verified patterns from the existing codebase and applied to Phase 04 context:
### Database Schema Refactor (Drizzle)
```typescript
// src/db/schema.ts (NEW projects table + updated FKs)
// Clients now has slug field (optional, unique)
export const clients = pgTable("clients", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
name: text("name").notNull(),
brand_name: text("brand_name").notNull(),
brief: text("brief").notNull(),
token: text("token").notNull().unique().$defaultFn(() => nanoid()),
slug: text("slug").unique(), // NEW — optional, unique, URL-safe
archived: boolean("archived").notNull().default(false),
created_at: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
});
// NEW projects table
export const projects = pgTable("projects", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
client_id: text("client_id").notNull().references(() => clients.id, { onDelete: "cascade" }),
name: text("name").notNull(), // brand/project name
accepted_total: numeric("accepted_total", { precision: 10, scale: 2 }).default("0"), // denormalized
archived: boolean("archived").notNull().default(false),
created_at: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
});
// Phases: FK now points to projects, not clients
export const phases = pgTable("phases", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
project_id: text("project_id").notNull().references(() => projects.id, { onDelete: "cascade" }), // CHANGED from client_id
title: text("title").notNull(),
sort_order: integer("sort_order").notNull().default(0),
status: text("status").notNull().default("upcoming"),
});
// Payments: FK now points to projects, not clients
export const payments = pgTable("payments", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
project_id: text("project_id").notNull().references(() => projects.id, { onDelete: "cascade" }), // CHANGED from client_id
label: text("label").notNull(),
amount: numeric("amount", { precision: 10, scale: 2 }).notNull(),
status: text("status").notNull().default("da_saldare"),
paid_at: timestamp("paid_at", { withTimezone: true }),
});
// Quote items: FK now points to projects, not clients
export const quote_items = pgTable("quote_items", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
project_id: text("project_id").notNull().references(() => projects.id, { onDelete: "cascade" }), // CHANGED from client_id
service_id: text("service_id").references(() => service_catalog.id, { onDelete: "restrict" }),
quantity: numeric("quantity", { precision: 10, scale: 2 }).notNull(),
unit_price: numeric("unit_price", { precision: 10, scale: 2 }).notNull(),
subtotal: numeric("subtotal", { precision: 10, scale: 2 }).notNull(),
custom_label: text("custom_label"),
});
// Time entries: FK now points to projects, not clients
export const time_entries = pgTable("time_entries", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
project_id: text("project_id").notNull().references(() => projects.id, { onDelete: "cascade" }), // CHANGED from client_id
started_at: timestamp("started_at", { withTimezone: true }).notNull().defaultNow(),
ended_at: timestamp("ended_at", { withTimezone: true }),
duration_seconds: integer("duration_seconds"),
});
// Documents: FK now points to projects, not clients
export const documents = pgTable("documents", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
project_id: text("project_id").notNull().references(() => projects.id, { onDelete: "cascade" }), // CHANGED from client_id
label: text("label").notNull(),
url: text("url").notNull(),
created_at: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
});
// Notes: FK now points to projects, not clients
export const notes = pgTable("notes", {
id: text("id").primaryKey().$defaultFn(() => nanoid()),
project_id: text("project_id").notNull().references(() => projects.id, { onDelete: "cascade" }), // CHANGED from client_id
body: text("body").notNull(),
created_at: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
});
// NEW settings table for global admin settings (e.g., target hourly rate)
export const settings = pgTable("settings", {
key: text("key").primaryKey(),
value: text("value").notNull(),
updated_at: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow(),
});
// Relations updated
export const projectsRelations = relations(projects, ({ one, many }) => ({
client: one(clients, { fields: [projects.client_id], references: [clients.id] }),
phases: many(phases),
payments: many(payments),
documents: many(documents),
notes: many(notes),
quote_items: many(quote_items),
}));
```
### Admin Query Layer — getProjectFullDetail
```typescript
// src/lib/admin-queries.ts (NEW function, following getClientFullDetail pattern)
export type ProjectFullDetail = {
project: Project & { client: Client };
phases: Array<Phase & { tasks: Array<Task & { deliverables: Deliverable[] }> }>;
payments: Payment[];
documents: Document[];
notes: Note[];
comments: Comment[];
quoteItems: QuoteItemWithLabel[];
activeServices: ServiceCatalog[];
totalTrackedSeconds: number; // for profitability calc
};
export async function getProjectFullDetail(id: string): Promise<ProjectFullDetail | null> {
const projectRows = await db
.select()
.from(projects)
.where(eq(projects.id, id))
.limit(1);
if (projectRows.length === 0) return null;
const project = projectRows[0];
// Fetch parent client
const clientRows = await db
.select()
.from(clients)
.where(eq(clients.id, project.client_id))
.limit(1);
const client = clientRows[0] || null;
// Fetch all phases for this PROJECT (not client)
const phasesRows = await db
.select()
.from(phases)
.where(eq(phases.project_id, id))
.orderBy(asc(phases.sort_order));
const phaseIds = phasesRows.map((p) => p.id);
// Fetch tasks scoped to this project's phases
const tasksRows = phaseIds.length === 0
? []
: await db
.select()
.from(tasks)
.where(inArray(tasks.phase_id, phaseIds))
.orderBy(asc(tasks.sort_order));
const taskIds = tasksRows.map((t) => t.id);
// Fetch deliverables scoped to this project's tasks
const deliverablesRows = taskIds.length === 0
? []
: await db
.select()
.from(deliverables)
.where(inArray(deliverables.task_id, taskIds));
// Payments for this PROJECT (not client)
const paymentsRows = await db
.select()
.from(payments)
.where(eq(payments.project_id, id));
// Documents for this PROJECT (not client)
const documentsRows = await db
.select()
.from(documents)
.where(eq(documents.project_id, id))
.orderBy(asc(documents.created_at));
// Notes for this PROJECT (not client)
const notesRows = await db
.select()
.from(notes)
.where(eq(notes.project_id, id))
.orderBy(asc(notes.created_at));
// Comments (polymorphic on entity_id) — collect all tasks, deliverables, and the project itself
const allEntityIds = [id, ...taskIds, ...deliverablesRows.map((d) => d.id)];
const commentsRows = allEntityIds.length === 0
? []
: await db
.select()
.from(comments)
.where(inArray(comments.entity_id, allEntityIds))
.orderBy(asc(comments.created_at));
// Quote items for this PROJECT (not client)
const quoteItemRows: QuoteItemWithLabel[] = await db
.select({
id: quote_items.id,
label: sql<string>`COALESCE(${service_catalog.name}, ${quote_items.custom_label})`,
custom_label: quote_items.custom_label,
service_id: quote_items.service_id,
quantity: quote_items.quantity,
unit_price: quote_items.unit_price,
subtotal: quote_items.subtotal,
})
.from(quote_items)
.leftJoin(service_catalog, eq(quote_items.service_id, service_catalog.id))
.where(eq(quote_items.project_id, id))
.orderBy(asc(quote_items.id));
// Active services (unchanged)
const activeServiceRows = await db
.select()
.from(service_catalog)
.where(eq(service_catalog.active, true))
.orderBy(asc(service_catalog.name));
// Total tracked seconds for this PROJECT (for profitability calc)
const totalRes = await db
.select({
total: sql<string>`coalesce(sum(${time_entries.duration_seconds}), 0)`,
})
.from(time_entries)
.where(eq(time_entries.project_id, id));
const totalTrackedSeconds = totalRes[0] ? parseInt(totalRes[0].total) : 0;
// Rebuild hierarchy
const phasesWithTasks = phasesRows.map((phase) => ({
...phase,
tasks: tasksRows
.filter((t) => t.phase_id === phase.id)
.map((task) => ({
...task,
deliverables: deliverablesRows.filter((d) => d.task_id === task.id),
})),
}));
return {
project: { ...project, client } as any,
phases: phasesWithTasks,
payments: paymentsRows,
documents: documentsRows,
notes: notesRows,
comments: commentsRows,
quoteItems: quoteItemRows,
activeServices: activeServiceRows,
totalTrackedSeconds,
};
}
```
### Slug Resolution in Middleware
```typescript
// src/proxy.ts (UPDATED for slug-first resolution)
export async function proxy(request: NextRequest) {
const pathname = request.nextUrl.pathname;
// ── ADMIN GUARD ──────────────────────────────────────────────────────────
if (pathname.startsWith("/admin")) {
if (pathname === "/admin/login" || pathname.startsWith("/api/auth")) {
return NextResponse.next();
}
const token = await getToken({
req: request,
secret: process.env.NEXTAUTH_SECRET,
});
if (!token) {
const loginUrl = new URL("/admin/login", request.url);
loginUrl.searchParams.set("callbackUrl", pathname);
return NextResponse.redirect(loginUrl);
}
return NextResponse.next();
}
// ── CLIENT TOKEN/SLUG GUARD ─────────────────────────────────────────────
if (pathname.startsWith("/c/")) {
const slugOrTokenMatch = pathname.match(/^\/c\/([a-zA-Z0-9_-]+)/);
if (!slugOrTokenMatch) {
return NextResponse.rewrite(new URL("/not-found", request.url));
}
const slugOrToken = slugOrTokenMatch[1];
try {
// TRY SLUG FIRST — call internal API to resolve slug → client
const validateUrl = new URL(
`/api/internal/validate-slug?slug=${encodeURIComponent(slugOrToken)}`,
request.url
);
let res = await fetch(validateUrl.toString());
// If slug not found, fall back to TOKEN validation (existing pattern)
if (!res.ok) {
const validateTokenUrl = new URL(
`/api/internal/validate-token?token=${encodeURIComponent(slugOrToken)}`,
request.url
);
res = await fetch(validateTokenUrl.toString());
}
if (!res.ok) {
return NextResponse.rewrite(new URL("/not-found", request.url));
}
return NextResponse.next();
} catch {
return NextResponse.rewrite(new URL("/not-found", request.url));
}
}
return NextResponse.next();
}
export const config = {
matcher: ["/admin/:path*", "/c/:path*"],
};
```
### New Internal API Route for Slug Validation
```typescript
// src/app/api/internal/validate-slug/route.ts (NEW)
import { NextRequest, NextResponse } from "next/server";
import { db } from "@/db";
import { clients } from "@/db/schema";
import { eq } from "drizzle-orm";
export async function GET(request: NextRequest) {
const slug = request.nextUrl.searchParams.get("slug");
if (!slug) {
return NextResponse.json({ error: "slug required" }, { status: 400 });
}
const rows = await db
.select({ id: clients.id })
.from(clients)
.where(eq(clients.slug, slug))
.limit(1);
if (rows.length === 0) {
return NextResponse.json({ error: "not found" }, { status: 404 });
}
return NextResponse.json({ clientId: rows[0].id }, { status: 200 });
}
```
### Timer Actions Refactored for Project Scope
```typescript
// src/app/admin/timer-actions.ts (UPDATED for project_id)
"use server";
import { revalidatePath } from "next/cache";
import { db } from "@/db";
import { time_entries } from "@/db/schema";
import { eq, isNull } from "drizzle-orm";
import { nanoid } from "nanoid";
export async function startTimer(projectId: string): Promise<{ entryId: string }> {
// Stop any currently running session (still global: only one timer active per admin)
const running = await db
.select({ id: time_entries.id })
.from(time_entries)
.where(isNull(time_entries.ended_at));
for (const r of running) {
const now = new Date();
const entry = await db
.select({ started_at: time_entries.started_at })
.from(time_entries)
.where(eq(time_entries.id, r.id))
.limit(1);
if (entry[0]) {
const secs = Math.round((now.getTime() - new Date(entry[0].started_at).getTime()) / 1000);
await db
.update(time_entries)
.set({ ended_at: now, duration_seconds: secs })
.where(eq(time_entries.id, r.id));
}
}
// Create new entry scoped to PROJECT (not client)
const id = nanoid();
await db.insert(time_entries).values({ id, project_id: projectId });
revalidatePath("/admin");
return { entryId: id };
}
export async function stopTimer(entryId: string): Promise<void> {
const rows = await db
.select({ started_at: time_entries.started_at })
.from(time_entries)
.where(eq(time_entries.id, entryId))
.limit(1);
if (!rows[0]) return;
const now = new Date();
const secs = Math.round((now.getTime() - new Date(rows[0].started_at).getTime()) / 1000);
await db
.update(time_entries)
.set({ ended_at: now, duration_seconds: secs })
.where(eq(time_entries.id, entryId));
revalidatePath("/admin");
}
```
### Profitability Analytics Card (Component Pattern)
```typescript
// src/components/admin/ProfitabilityCard.tsx (NEW)
import { Project } from "@/db/schema";
export function ProfitabilityCard({
project,
totalTrackedSeconds,
targetHourlyRate,
}: {
project: Project & { accepted_total: string };
totalTrackedSeconds: number;
targetHourlyRate: number; // e.g., 50 €/h
}) {
const hours = totalTrackedSeconds / 3600;
const acceptedTotal = parseFloat(project.accepted_total || "0");
// €/h real = accepted_total ÷ hours
const realHourlyRate = hours > 0 ? acceptedTotal / hours : 0;
// Ideal cost = target_rate × hours
const idealCost = targetHourlyRate * hours;
// Delta = profit/loss
const delta = acceptedTotal - idealCost;
const deltaIsProfit = delta >= 0;
return (
<div className="bg-white rounded-lg border border-[#e5e7eb] p-4 space-y-3">
<h3 className="font-medium text-[#1a1a1a]">Profittabilità</h3>
<div className="grid grid-cols-2 gap-3 text-sm">
<div>
<p className="text-[#71717a] text-xs">Ore lavorate</p>
<p className="font-mono font-semibold text-[#1a1a1a]">{hours.toFixed(1)}h</p>
</div>
<div>
<p className="text-[#71717a] text-xs">Importo accettato</p>
<p className="font-mono font-semibold text-[#1a1a1a]">€{acceptedTotal.toFixed(2)}</p>
</div>
</div>
<div className="border-t border-[#f4f4f5] pt-3 space-y-2 text-sm">
<div className="flex justify-between">
<span className="text-[#71717a]">€/h reale</span>
<span className="font-mono font-semibold text-[#1a1a1a]">€{realHourlyRate.toFixed(2)}/h</span>
</div>
<div className="flex justify-between">
<span className="text-[#71717a]">€/h target</span>
<span className="font-mono font-semibold text-[#71717a]">€{targetHourlyRate.toFixed(2)}/h</span>
</div>
<div className="flex justify-between">
<span className="text-[#71717a]">Costo ideale</span>
<span className="font-mono font-semibold text-[#1a1a1a]">€{idealCost.toFixed(2)}</span>
</div>
</div>
<div className={`border-t border-[#f4f4f5] pt-3 flex justify-between items-center`}>
<span className="text-[#71717a]">Delta (guadagno/perdita)</span>
<span className={`font-mono font-bold ${deltaIsProfit ? "text-green-600" : "text-red-600"}`}>
{deltaIsProfit ? "+" : ""}€{delta.toFixed(2)}
</span>
</div>
</div>
);
}
```
---
## State of the Art
| Old Approach | Current Approach | When Changed | Impact |
|--------------|------------------|--------------|--------|
| Single project per client | Multi-project per client | Phase 04 | Clients can now manage multiple independent brands/projects with separate workspaces and profitability tracking |
| Client as primary work container | Project as primary work container | Phase 04 | Admin workspace structure mirrors project, not client. Client API queries projects, not client directly |
| Timer at client level | Timer at project level | Phase 04 | Hours tracked independently per project, enabling per-project profitability analysis |
| Token-only client links | Token + slug client links | Phase 04 | More user-friendly URLs (e.g., /c/mario-rossi instead of /c/xyzabc123). Token remains as fallback. |
| Hardcoded profitability target | Global settings table for target rate | Phase 04 | Admin can adjust target hourly rate from UI without changing code. Flexible for future settings. |
**Deprecated/outdated:**
- Client-level `accepted_total` field remains in schema for backward compat but becomes unused; project-level accepted_total is the source of truth.
- Old client detail workspace layout is cloned for project detail; both exist but admin only uses project workspace going forward.
---
## Assumptions Log
| # | Claim | Section | Risk if Wrong |
|---|-------|---------|---------------|
| A1 | "Only one timer active globally (per admin)" is the design intent, not per-project independence | Locked Decision D-15, Timer Pitfall | If per-project timers are expected, timer-actions refactor is wrong; need concurrent timer support instead of auto-stop logic. Clarify with user before implementing. |
| A2 | Hard migration (drop/recreate tables) is acceptable because all data is test data | Runtime State Inventory | If there are production customer records, hard migration will cause data loss. Verify no production data exists before schema push. |
| A3 | `settings` table with key-value structure is acceptable over env vars | Claude's Discretion | If user later requires non-DB storage for settings (e.g., Redis cache, config file), table approach is still compatible; no blocking constraint. |
| A4 | Slug-first middleware resolution (slug lookup before token fallback) is the intended order | Locked Decision D-06, D-08 | If token validation should be checked first (for legacy reasons), middleware order is reversed. Test both slug and token paths after implementation. |
| A5 | `comments` table remains polymorphic (entity_id) and does NOT move to project_id | Locked Decision D-02 | If comments should be scoped per-project (unlikely), add project_id FK and update all comment queries. Currently comments are global per entity, which is correct. |
**If this table is empty:** Not applicable — all claims verified against CONTEXT.md locked decisions.
---
## Open Questions
1. **Profitability analytics default target rate?**
- What we know: User mentioned 50€/h as an example; needs global setting.
- What's unclear: Should there be a fallback default (e.g., 50€/h) if settings table is empty, or should admin be forced to set it?
- Recommendation: Initialize settings table with `target_hourly_rate = '50.00'` as default during first project workspace load. Admin can override from /admin/impostazioni.
2. **Multi-project client dashboard routing — how should it work?**
- What we know: 1 project = direct view, 2+ projects = tabs.
- What's unclear: Should the URL path for client dashboard change? Stay `/c/[token]` for all projects, or add `/c/[token]/projects/[id]`?
- Recommendation: Keep URL `/c/[token]` (or `/c/[slug]`), let server-side logic choose whether to render single project view or tabs. Tabs can have internal navigation (e.g., URL hash or search param) to switch between projects without page reload.
3. **Analytics page aggregation scope?**
- What we know: D-22 says "Statistiche page shows aggregated profitability for all projects + breakdown per client."
- What's unclear: Should /admin/analytics show global profitability (sum of all projects for all clients) or be filterable by client/date range?
- Recommendation: Start simple: global profitability table with columns: Client, Projects, Total Hours, Total Revenue, Avg €/h, Profit/Loss. Filter by client optional (defer to Phase 4.1 if needed).
4. **Project archival behavior?**
- What we know: `projects.archived` field exists; D-13 doesn't mention archival UI.
- What's unclear: Should archived projects be hidden from /admin/projects list or filtered to separate tab?
- Recommendation: Hide archived projects by default (like clients list), add "Mostra archiviati" toggle link. Archival doesn't delete data, just hides it.
---
## Environment Availability
(Phase 04 is code/DB changes only — no external dependencies.)
| Dependency | Required By | Available | Version | Fallback |
|------------|------------|-----------|---------|----------|
| Neon Postgres | Schema migration + queries | ✓ | Active (from Phase 1) | — |
| Next.js API routes | Slug validation route | ✓ | 16 (installed) | — |
| Drizzle ORM | Schema migration + query building | ✓ | Current (installed) | — |
| Auth.js v4 | Admin session check (existing) | ✓ | Current (installed) | — |
| shadcn/ui Tabs | Multi-project dashboard tabs | ✓ | Current (installed) | Could fall back to native `<select>` dropdown, but Tabs is already in use |
**Missing dependencies with no fallback:** None.
**Missing dependencies with fallback:** Tabs component could be replaced with a `<select>` dropdown if shadcn/ui is ever removed, but this is a UI detail, not a blocker.
---
## Sources
### Primary (HIGH confidence)
- **CONTEXT.md** (Phase 04 decisions) — All 22 locked decisions directly from user's discuss phase. D-01 through D-22, Claude's Discretion, Deferred Ideas sections.
- **Codebase inspection** — Verified existing schema (schema.ts), admin queries (admin-queries.ts), middleware pattern (proxy.ts), timer actions (timer-actions.ts), client view pattern (client-view.ts), admin workspace layout (clients/[id]/page.tsx).
- **REQUIREMENTS.md** — PROJ-01 through PROJ-05 mapped to implementation guidance.
- **ROADMAP.md** — Phase 04 goal and success criteria verified.
### Secondary (MEDIUM confidence)
- **Next.js 16 App Router patterns** — Edge middleware, server actions, API routes all verified against existing project structure.
- **Drizzle ORM query patterns** — Relations, WHERE scoping, parallel queries all verified against Phase 13 implementation (getAllClientsWithPayments, getClientFullDetail, timer-actions).
- **shadcn/ui Tabs component** — Already in use in admin workspace (clients/[id]/page.tsx); no additional research needed.
### Tertiary (LOW confidence)
- None — all findings tied to locked decisions and verified codebase patterns.
---
## Metadata
**Confidence breakdown:**
- **Standard stack:** HIGH — all libraries already installed and used; no new dependencies.
- **Architecture:** HIGH — locked decisions in CONTEXT.md eliminate discretion; patterns clone from existing workspaces.
- **Pitfalls:** HIGH — identified from common FK migration mistakes, middleware routing, query scoping issues observed in similar refactors.
- **Environment:** HIGH — no external dependencies; Neon, Next.js, Drizzle all active and verified.
**Research date:** 2026-05-21
**Valid until:** 2026-06-04 (14 days — architecture stable, no fast-moving libraries)
**Next phase:** `/gsd-plan-phase 04` will create 45 plans for vertical-slice execution (Schema Wave 0 → Core Routing → Admin UI → Client UI + Analytics).
+191
View File
@@ -0,0 +1,191 @@
# Architecture — ClientHub Freelancer Client Portal
**Project:** ClientHub (welcomeclient.iamcavalli.net)
**Researched:** 2026-05-09
**Confidence:** HIGH
---
## Component Boundaries
Single Next.js application on Vercel. No separate backend. All server logic lives in Route Handlers (`/api/**`). One Postgres database (Neon serverless) accessed via Drizzle ORM. Admin auth via env-var secret + cookie. Client access via UUID token in URL — no auth library needed for clients.
| Component | Responsibility | Communicates With |
|-----------|---------------|-------------------|
| Client Portal `/c/[token]` | Read-only view: status, phases, tasks, deliverables, payments, documents | API Routes (GET only) |
| Admin Dashboard `/admin` | List all clients with status summary | API Routes (full CRUD) |
| Admin Client Workspace `/admin/clients/[id]` | Edit phases, tasks, deliverables, payments, documents | API Routes (full CRUD) |
| Service Catalog Manager `/admin/catalog` | CRUD on service items + unit prices | API Routes (catalog entity) |
| Quote Builder `/admin/clients/[id]/quote` | Compose quote from catalog items, write `accepted_total` to client row | Catalog + API Routes |
| Comments System | Client posts on task/deliverable; admin replies | API Route POST |
| Approval Flow | Client PATCHes a deliverable to `approved` | API Route, validates token ownership |
| API Routes `/api/**` | Validate token or admin session; query/mutate DB; return JSON | Postgres only |
| Database | Single source of truth | API Routes only — never queried from browser |
---
## Data Flow
**Client reading their dashboard:**
```
Browser → GET /c/[token]
→ Next.js server component
→ DB: clients WHERE token = [token] → 404 if missing
→ JOIN: project + phases + tasks + deliverables + payments + documents
→ Omit: quote_items, service prices
→ Render read-only portal
```
**Client posting a comment:**
```
Browser → POST /api/comments { token, entity_type, entity_id, body }
→ Validate token → write comment { author: 'client' }
→ 201 → re-fetch thread
```
**Client approving a deliverable:**
```
Browser → PATCH /api/deliverables/[id]/approve { token }
→ Validate token owns deliverable → set status='approved', approved_at=now()
→ Return updated deliverable
```
**Admin editing:**
```
Browser (admin) → PATCH /api/admin/tasks/[id] + admin cookie
→ Validate session → update row → return updated task
```
**Quote building:**
```
Admin UI selects services → computes line items
→ POST /api/admin/clients/[id]/quote { line_items[], accepted_total }
→ Write quote_items rows + write clients.accepted_total (denormalized)
→ Client portal reads clients.accepted_total — never touches quote_items
```
---
## Data Model
```
clients
id UUID PK
name TEXT
brand_name TEXT
brief TEXT
token UUID UNIQUE ← the secret link key (separate from PK!)
accepted_total NUMERIC ← denormalized; only price client ever sees
created_at TIMESTAMPTZ
phases
id UUID PK
client_id UUID → clients.id
title TEXT
sort_order INT
status TEXT (upcoming | active | done)
tasks
id UUID PK
phase_id UUID → phases.id
title TEXT
description TEXT
status TEXT (todo | in_progress | done)
sort_order INT
deliverables
id UUID PK
task_id UUID → tasks.id
title TEXT
url TEXT ← external link (Google Drive, PDF, etc.)
status TEXT (pending | submitted | approved)
approved_at TIMESTAMPTZ ← immutable audit trail
comments
id UUID PK
entity_type TEXT (task | deliverable)
entity_id UUID
author TEXT (client | admin)
body TEXT
created_at TIMESTAMPTZ
payments
id UUID PK
client_id UUID → clients.id
label TEXT ("Acconto 50%" / "Saldo 50%")
amount NUMERIC
status TEXT (da_saldare | inviata | saldato)
paid_at TIMESTAMPTZ
documents
id UUID PK
client_id UUID → clients.id
label TEXT
url TEXT
created_at TIMESTAMPTZ
service_catalog
id UUID PK
name TEXT
description TEXT
unit_price NUMERIC
active BOOLEAN
quote_items
id UUID PK
client_id UUID → clients.id
service_id UUID → service_catalog.id
quantity NUMERIC
unit_price NUMERIC ← snapshot at time of quote
subtotal NUMERIC
-- NEVER exposed via client API
```
**Key design decisions:**
- `clients.token` is the only secret. Rotation = single UPDATE. No session store needed.
- `clients.accepted_total` is deliberately denormalized so client API never touches `quote_items`.
- Approval `approved_at` stored as immutable audit trail — disputes resolved by timestamp.
- `comments` use `entity_type + entity_id` polymorphic pair — correct at this scale.
- `payments` always two rows per client (created when quote is finalized).
---
## Suggested Build Order
```
1. DB schema + migrations
└─ everything depends on this
2. API: token lookup + project read (GET only)
└─ unblocks client portal
3. Client portal UI /c/[token]
└─ the core deliverable; clients need this first
4. Admin auth middleware (env-var secret, cookie check)
└─ gate before admin routes go live
5. Admin: client list + client workspace CRUD
└─ phases, tasks, status, documents, payments
6. Comments system + deliverable approval
└─ depends on both client portal and admin workspace
7. Service catalog CRUD ← can run parallel with step 5
└─ independent of client-facing features
8. Quote builder
└─ depends on catalog + client entity
9. Claude onboarding flow (v2)
└─ depends on all CRUD APIs being complete
```
---
## Roadmap Implications
- Phase 1: DB schema + token API + client portal (all three coupled)
- Phase 2: Admin auth + CRUD management + comments + approvals
- Phase 3: Service catalog + quote builder
- Phase 4 (v2): Claude onboarding flow
+107
View File
@@ -0,0 +1,107 @@
# FEATURES.md — ClientHub Freelancer Client Portal
**Domain:** Freelancer client portal — solo personal branding consultant
**Project:** ClientHub (welcomeclient.iamcavalli.net)
**Researched:** 2026-05-09
**Confidence:** HIGH
---
## Context
Two asymmetric roles. Admin (the freelancer) has full CRUD. Client (read + lightweight interaction) accesses via secret URL — no login, no account — and can view, comment, and approve. The product competes indirectly with Notion client portals, HoneyBook, Dubsado, and bespoke agency portals. The differentiator is zero-friction secret link access and personal brand positioning.
---
## Table Stakes
Features clients expect when opening any project portal. Missing these causes confusion, distrust, or support overhead.
| Feature | Why Expected | Complexity | Notes |
|---------|--------------|------------|-------|
| Project overview at a glance | Client needs to know "where we are" without reading walls of text | Low | Name, brand, brief, current phase |
| Phase + task status visibility | Primary client question is "what's done, what's next" | Low | Phases with nested tasks; status per task (todo / in progress / done) |
| Deliverable approval | Client must formally sign off on outputs | Medium | Per-deliverable approve action; state persists; admin sees approval timestamp |
| Inline commenting on tasks/deliverables | Feedback and questions without email | Medium | Flat comments sufficient for v1; threading is nice-to-have |
| Document / file links | Deliverables, briefs, contracts surface in the portal | Low | Links to Google Drive, PDF, external URL; no file hosting needed |
| Payment status visibility | Client needs to know what they owe | Low | Deposit 50% + balance 50%; three states each: pending / invoiced / paid |
| Total quoted amount (not itemized) | Client expects to see the agreed number | Low | Single total; line items are admin-only |
| Mobile-readable layout | Clients open links on phones | Low | Responsive web; no native app |
| Persistent secret link | Link must not expire or rotate without notice | Low | UUIDs in DB, never regenerated unless admin resets explicitly |
| Trustworthy, branded appearance | First impression determines confidence in the consultant | Low | Logo, brand colors, professional typography — not a generic SaaS look |
---
## Differentiators
Not expected, but meaningfully improve experience or workflow.
| Feature | Value Proposition | Complexity | Notes |
|---------|-------------------|------------|-------|
| Decision log / history | Running record of agreed decisions — eliminates "we never agreed on that" disputes | Low | Append-only note stream visible to client; admin writes entries |
| Phase progress indicator | Visual progress bar gives a sense of momentum | Low | Derived from task completion %; no extra data model needed |
| "Last updated" timestamp on dashboard | Shows the portal is live and maintained | Low | Trivially derived from DB updated_at |
| Admin overview: all clients at a glance | Freelancer scans all active projects and overdue payments in one view | Medium | List with status badges; payment alert if overdue |
| Payment status badge with clear labels | Color-coded states (red = unpaid, yellow = invoiced, green = paid) | Low | Client sees their own; admin sees all |
| Shareable link reset | Admin can invalidate and regenerate a client's link if it leaks | Low | DB field update + redirect; rarely used but reassuring |
| Service catalog | Admin builds quotes from a curated menu of services; reusable across clients | Medium | Lookup table; admin-only; used by Claude in v2 |
| Claude-assisted onboarding (v2) | Generates phases + quote draft from a brief — massively speeds up admin work | High | Explicitly v2 in PROJECT.md |
---
## Anti-Features
Deliberately NOT building these.
| Anti-Feature | Why Avoid | What to Do Instead |
|--------------|-----------|-------------------|
| Client login / account creation | Adds friction with no benefit for a small client list | Secret UUID link |
| In-app invoicing / PDF generation | Accounting is out of scope | Show payment status only |
| File upload / storage | Massive complexity | Link to Google Drive or Dropbox |
| Email / SMS notifications | Transactional email infrastructure is heavy | Manual communication fine for small client list |
| Multi-admin / team roles | Freelancer works alone | Single admin |
| Client-editable project structure | Clients editing phases corrupts admin's source of truth | Comment and approve only |
| Itemized pricing visible to client | Erodes commercial confidentiality | Single total; detail is admin-only |
| Kanban / drag-and-drop board | Phases are sequential, not a fluid backlog | Ordered phase list |
| Time tracking | Out of scope for project-based billing | Not relevant |
| Multi-language / i18n | Single consultant, single-market | Hardcode interface language |
---
## Feature Dependencies
```
Secret link (UUID) → Client dashboard
Client dashboard → Phase/task display
Phase/task display → Deliverable approval
Phase/task display → Inline commenting
Admin client management → Secret link generation
Admin client management → Payment tracking
Service catalog → Quote building (admin picks from catalog)
Quote building → Payment tracking (total = basis for deposit/balance)
Service catalog → Claude onboarding v2
```
**Key insight:** Admin must create data before the client dashboard shows anything meaningful. Admin-first, then client.
---
## MVP Build Order
1. Admin: create/edit client record with secret link generation
2. Admin: create/edit phases and tasks per client
3. Admin: set payment amounts and statuses
4. Client dashboard: read-only view (overview, phases, tasks, payment status, documents)
5. Client: deliverable approval
6. Client: inline comments
7. Admin: all-clients overview
8. Admin: service catalog
9. v2: Claude-assisted onboarding
---
## Open Questions
- What happens when a client accidentally shares their secret link? Is link reset sufficient, or should there be an access log?
- Does the decision log need to be visible to clients from day one, or deferred?
- Should approval actions be reversible (un-approve)?
+177
View File
@@ -0,0 +1,177 @@
# Pitfalls — ClientHub Freelancer Client Portal
**Domain:** Freelancer client portal with secret-link access, solo developer, Vercel deploy
**Researched:** 2026-05-09
**Confidence:** HIGH
---
## Critical Pitfalls
### 1. Secret Links That Are Guessable or Enumerable
**What goes wrong:** If client tokens are generated from names, sequential IDs, or short strings, they can be enumerated. `/client/mario-rossi` or `/client/3` are not secrets.
**Prevention:**
- Generate tokens as cryptographically random UUIDs (v4) or nanoid (21 chars / ~126 bits of entropy)
- Never derive from name, company, or sequential ID
- Never log or display full tokens in admin analytics
**Warning signs:** Client slug contains the client's name. Token under 20 characters. Link can be reconstructed from info the client already knows.
**Phase mapping:** Phase 1 — data model + routing. Cannot be retrofitted after links are distributed.
---
### 2. Client API Exposes Admin Data (Hidden in UI Only)
**What goes wrong:** Developers fetch all client data and conditionally render fields. A technical client opens DevTools and sees the full quote breakdown — the exact thing the product prevents.
**Prevention:**
- Define two distinct data shapes: `ClientView` and `AdminView`
- Client API routes (`/api/c/[token]/`) return `ClientView` only — enforced server-side, not in the UI
- `accepted_total` goes on the `clients` row. Client API never queries `quote_items`
**Warning signs:** Client API response includes `lineItems` filtered in the frontend. "Client sees only total" enforced with `display: none`.
**Phase mapping:** Data model and API shape — Phase 1.
---
### 3. Data Loss from Vercel Filesystem / In-Memory Storage
**What goes wrong:** Vercel serverless functions are stateless. File writes to local filesystem and in-memory state are lost between invocations. SQLite on disk silently vanishes after cold start. Bug only manifests in production.
**Prevention:**
- External persistent DB from day one: Neon (Postgres free tier)
- Never write to `fs` for persistent data on Vercel
- Validate persistence in the first production deploy
**Warning signs:** Project stores data in a `data/` JSON directory. SQLite without a remote adapter. Data changes in one request not visible in the next.
**Phase mapping:** Day-one infrastructure decision — Phase 1, before any real data is entered.
---
### 4. Over-Engineering Before the First Client Uses It
**What goes wrong:** Building the Claude onboarding flow and quote generator before a single client has opened their dashboard. Real clients still managed via email while the portal is "almost ready."
**Prevention:**
- Hard success criterion for Phase 1: one client link is shareable and works
- Phase 1 ships read-only client dashboard only
- Collect direct feedback from one real client before adding features
**Warning signs:** >2 weeks pass without a shareable client link. Claude integration started before admin edit UI exists.
**Phase mapping:** Enforced by roadmap phase ordering.
---
### 5. Token Is the Primary Key (Unrotatable)
**What goes wrong:** A client forwards their link. The link appears in a screenshot. There is no mechanism to rotate without losing data or breaking bookmarks. Worse if token = primary key: rotation requires a migration.
**Prevention:**
- Data model: stable internal UUID as PK; secret token is a separate, independently updatable field
- Build "Regenerate link" in admin area during Phase 2 — it's a single field UPDATE
- Overwriting the token field is sufficient to invalidate the old link
**Warning signs:** Token is used as PK of the client record. No admin affordance to regenerate a link.
**Phase mapping:** Data model separation — Phase 1. Admin rotation UI — Phase 2.
---
### 6. Client Approval Has No Immutable Record
**What goes wrong:** Client clicks "Approve." Later: "I never approved that." If approval is a boolean with no timestamp, there is no evidence. Weakens your position in commercial disputes.
**Prevention:**
- Store `approved_at` timestamp alongside the approval boolean — from day one
- Display approval timestamp in admin view
- Approvals are immutable for clients — no "undo" button
**Warning signs:** Approval is a boolean column with no timestamp. Client-visible "undo approval" button exists.
**Phase mapping:** Schema — Phase 1. Display in admin — Phase 2.
---
## Moderate Pitfalls
### 7. Payment Status Without a Valid State Machine
Payment has three states (da_saldare / inviata / saldato) for two payments. If transitions are not enforced, the dashboard shows contradictory states.
**Prevention:** Enforce valid transitions: `da_saldare → inviata → saldato`. Admin UI offers only valid next states.
**Phase mapping:** Admin UI — Phase 2.
---
### 8. Google Drive Links That Rot
Document links break when sharing settings change or the Drive account changes. Client sees a broken link to their own deliverable.
**Prevention:** Store a display name alongside the URL so broken links are visible in admin. Build a simple "update link" affordance in Phase 2.
---
### 9. Admin Area With No Real Access Control
`/admin` is a secret route with no authentication. A client who guesses the URL accesses all client data.
**Prevention:** Add Next.js middleware check against `ADMIN_PASSWORD` env variable before Phase 2 ships. Never rely on security through obscurity for a route that contains all client data.
**Phase mapping:** Phase 2, before admin area contains real data.
---
### 10. Comments Scope Creeping Into Threading
"Clients can leave comments" becomes complex when replies, read/unread state, and notifications are added. Can double Phase 1 scope.
**Prevention:** Phase 1 ships comments as a flat append-only list per task. No threading, no replies, no email notifications.
---
## Minor Pitfalls
### 11. DNS Configuration as a Last-Minute Task
`welcomeclient.iamcavalli.net` requires a CNAME to Vercel's DNS. Propagation takes minutes to hours. Doing this the day of a client demo misses the deadline.
**Prevention:** Configure DNS in Phase 1 before the UI is complete. Verify propagation independently.
---
### 12. Mobile Responsiveness as an Afterthought
Clients open the link on their phone from a shared message. A broken mobile layout is the first impression.
**Prevention:** Use Tailwind mobile-first defaults from the start. Test on a real phone before any link is sent.
---
### 13. No Empty States
A new client record with no tasks shows a blank page. The client assumes something is broken.
**Prevention:** Design minimal empty states for no-tasks, no-documents, no-comments.
---
## Phase-Specific Warnings Summary
| Phase Topic | Likely Pitfall | Mitigation |
|---|---|---|
| Client token generation | Guessable slug from name | Crypto-random UUID/nanoid, never name-derived |
| Client-facing API | Admin data in JSON response | `ClientView` type enforced server-side |
| Storage choice | Vercel filesystem not persistent | External DB (Neon) before first data write |
| Admin area access | No real auth | Middleware check before Phase 2 ships |
| Approval recording | Boolean-only, no audit trail | Store `approved_at` from day one |
| Token in data model | Token = PK, unrotatable | Separate stable ID from rotatable token field |
| Phase ordering | Claude flow before dashboard | Enforce: client view → admin edit → Claude |
| Comments | Threading scope creep | Flat list in Phase 1 |
| DNS | Last-minute failure | Configure and verify in Phase 1 |
+132
View File
@@ -0,0 +1,132 @@
# Technology Stack — ClientHub Freelancer Client Portal
**Project:** ClientHub (welcomeclient.iamcavalli.net)
**Researched:** 2026-05-09
**Confidence:** HIGH
---
## Recommended Stack
### Core Framework
| Technology | Version | Purpose | Why |
|------------|---------|---------|-----|
| Next.js | 15.x (latest stable) | Full-stack app framework | App Router + Server Actions replace a separate API layer. Vercel-native: no adapter needed. First-class TypeScript. |
| React | 19.x | UI rendering | Bundled with Next.js. Server Components eliminate client-side waterfalls for the read-heavy client portal. |
| TypeScript | 5.x | Type safety | Drizzle + Zod give end-to-end type inference from DB schema to form validation. |
**Why NOT Remix / SvelteKit / Astro:** They work on Vercel but add unfamiliarity overhead with no gain at this scale.
---
### Database
| Technology | Purpose | Why |
|------------|---------|-----|
| Neon (serverless Postgres) | Primary database | Free plan: 0.5 GB storage + 100 CU-hours/month — sufficient for 520 clients. Scales to zero between uses. Native Vercel integration that auto-injects DATABASE_URL per preview branch. |
| Drizzle ORM | DB access + migrations | Lightest-weight TS ORM. Ships `drizzle-orm/neon-http` serverless driver — no persistent TCP connections, works in Vercel Node and Edge runtimes for free. Schema-as-code with `drizzle-kit` handles migrations. |
**Why NOT Prisma:** Needs PgBouncer or Prisma Accelerate ($) for serverless connection pooling. Drizzle's `neon-http` handles this natively at zero cost.
**Why NOT Supabase:** Adds RLS, Realtime, and Auth overhead you don't need and will have to maintain.
---
### Authentication
| Technology | Purpose | Why |
|------------|---------|-----|
| Auth.js (next-auth) v4 | Admin session management | Credentials provider with a single admin account. Session stored as signed JWT cookie. No user table in DB. |
| Next.js Middleware (custom) | Client secret-link validation | Each client has a `secretToken` (nanoid, 21 chars) stored in DB. Middleware reads `[token]` path segment, validates against Neon, returns 404 on miss. Runs at the edge before any page renders. |
| nanoid | Token generation | Cryptographically secure, URL-safe, 21-char default (~126 bits of entropy). Generated once at client creation. |
**Auth flow summary:**
- `/admin/*` → Auth.js session required (single admin account)
- `/c/[token]/*` → Middleware validates token against Neon, 404 on miss
- Client pages: zero auth library overhead
---
### UI
| Technology | Purpose | Why |
|------------|---------|-----|
| Tailwind CSS v4 | Utility-first styling | CSS-first configuration, zero runtime overhead. |
| shadcn/ui | Component library | Components copied into codebase (no runtime dep). Built on Radix UI (accessible). Provides: Badge, Progress, Card, Dialog, Table, Textarea, Select. |
| lucide-react | Icons | Tree-shaken, SVG-based, consistent. |
---
### Forms and Validation
| Technology | Purpose | Why |
|------------|---------|-----|
| Zod | Schema validation | Server-side in Server Actions + client-side with RHF resolver. Single source of truth for data shapes. |
| React Hook Form | Admin form state | Complex admin forms (client onboarding, task editing, quote builder). Client-facing pages use native `<form>` + Server Actions. |
---
### File Handling (v1)
None — document links stored as text fields in Postgres. Eliminates S3, CDN, and upload infrastructure from the initial build entirely.
**If direct uploads needed in v2:** UploadThing integrates directly with Next.js App Router, free tier (2 GB storage).
---
### Infrastructure
| Technology | Purpose | Why |
|------------|---------|-----|
| Vercel Hobby plan | Deploy + CDN + serverless | Native Next.js. Custom subdomain (`welcomeclient.iamcavalli.net`) via DNS CNAME. No Docker, VPS, or CI/CD to manage. |
| Neon Vercel Integration | DB branch per preview | Creates a fresh Neon branch per Git branch automatically. Safe schema migration testing. |
---
## Installation Sequence
```bash
# 1. Bootstrap Next.js
npx create-next-app@latest clienthub --typescript --tailwind --app --src-dir
# 2. Database
npm install drizzle-orm @neondatabase/serverless
npm install -D drizzle-kit
# 3. Auth
npm install next-auth
# 4. Token generation
npm install nanoid
# 5. Validation + Forms
npm install zod @hookform/resolvers react-hook-form
# 6. shadcn/ui
npx shadcn@latest init
npx shadcn@latest add badge button card dialog dropdown-menu input label progress select separator table textarea
```
---
## Key Architectural Decisions
1. **Secret-link without Auth.js:** Next.js Middleware validates `[token]` at the edge. Fast, zero client-side JS, 404 on invalid token.
2. **Server Actions for all mutations:** Task updates, comments, payment status — no REST API layer to maintain.
3. **Privacy model is a DB query filter:** Admin sees `quote_items`; clients see only `clients.accepted_total`. Not a UI filter — a DB design.
4. **Two auth systems, zero overlap:** Admin JWT cookie on `/admin/*`. Client token middleware on `/c/*`.
---
## Confidence Levels
| Area | Confidence | Notes |
|------|------------|-------|
| Next.js App Router | HIGH | Stable since Oct 2024 |
| Neon free tier | HIGH | 0.5 GB storage, 100 CU-hours/month |
| Drizzle + neon-http | HIGH | Free serverless driver, no connection pooling needed |
| Auth.js Credentials (admin) | HIGH | Mature, well-documented |
| nanoid secret tokens | HIGH | Cryptographically secure default |
| Tailwind v4 + Next.js | HIGH | Stable, PostCSS plugin verified |
| Vercel Hobby plan | HIGH | Custom subdomain supported |
+126
View File
@@ -0,0 +1,126 @@
# Project Research Summary
**Project:** ClientHub — welcomeclient.iamcavalli.net
**Domain:** Freelancer client portal (secret-link access, solo consultant)
**Researched:** 2026-05-09
**Confidence:** HIGH
---
## Executive Summary
ClientHub è un portale web a due ruoli per un consulente di personal branding. I clienti accedono via UUID segreto casuale — nessun account, nessun login, zero attrito. L'admin gestisce tutto: crea clienti, fasi, task, deliverable, pagamenti e preventivi. Il consensus della ricerca è chiaro: costruisci prima la dashboard cliente, poi l'admin CRUD, poi catalogo servizi e preventivi, poi il flusso Claude AI (v2).
Stack confermato: **Next.js 15 + Neon (Postgres) + Drizzle ORM + Auth.js + nanoid + Tailwind v4 + shadcn/ui**. Ogni scelta è ottimizzata per un developer solo su Vercel: nessun backend da mantenere, nessun costo di connection pooling, nessuna infrastruttura di upload file, nessuna libreria di auth per i clienti. Il meccanismo "secret link" è un Next.js Middleware edge check — veloce, zero client JS, 404 se il token non esiste.
I rischi dominanti sono architetturali, non tecnici. Se il token è la primary key diventa non-rotazionabile. Se la client API restituisce `quote_items` (anche nascosti nell'UI), un cliente con DevTools vede i prezzi dei singoli servizi. Se il progetto parte dal flusso Claude prima che un cliente possa aprire la sua dashboard, il portale non esce. Tutti e tre prevenibili con le decisioni corrette sul data model dal giorno uno.
---
## Key Findings
### Stack Raccomandato
| Tecnologia | Ruolo | Perché |
|------------|-------|--------|
| Next.js 15 (App Router) | Framework full-stack | Server Actions sostituiscono un'API REST separata; nativo Vercel |
| Neon (serverless Postgres) | Database principale | Free tier (0.5 GB, 100 CU-h/mese) sufficiente per 520 clienti; scala a zero |
| Drizzle ORM + neon-http | Accesso DB + migrazioni | Nessun costo di connection pooling; schema-as-code; inferenza TypeScript end-to-end |
| Auth.js v4 (Credentials) | Sessione admin | Account singolo, cookie JWT firmato, nessuna tabella utenti in DB |
| nanoid | Generazione token | 21 char, ~126 bit di entropia, URL-safe, crittograficamente sicuro |
| Tailwind v4 + shadcn/ui | UI | Componenti copiati nel codebase, accessibilità Radix UI, zero runtime dep |
| Zod + React Hook Form | Validazione e form | Schema unico; RHF solo per form admin complessi |
File upload deliberatamente esclusi dalla v1. I link ai documenti sono campi testo che puntano a Google Drive.
### Features v1
**Table stakes (obbligatori):**
- Panoramica progetto (nome, brand, brief, fase corrente)
- Visibilità fasi e task con stato (todo / in corso / fatto)
- Approvazione deliverable con timestamp immutabile
- Commenti inline su task e deliverable (lista piatta, no threading)
- Link a documenti esterni (solo URL, no file hosting)
- Stato pagamenti: acconto 50% + saldo 50% (da saldare / inviata / saldato)
- Totale preventivo accettato visibile al cliente (cifra unica, mai dettaglio)
- Layout mobile-ready
- Link segreto persistente e non-scadente
**Differenziatori (low-effort, includibili in v1):**
- Log decisioni / storico (nota append-only)
- Indicatore di avanzamento fase (derivato da % task completati)
- Timestamp "ultimo aggiornamento" sulla dashboard
- Vista admin: tutti i clienti con badge stato pagamenti
- Reset link segreto (single UPDATE, solo admin)
**Anti-features (mai costruire):** login cliente, PDF fatture in-app, multi-admin, struttura progetto modificabile dal cliente, prezzi singoli visibili al cliente, kanban board.
### Architettura
Singola applicazione Next.js su Vercel, un database Neon Postgres. Nessun backend separato.
**Due path di accesso isolati:**
- `/c/[token]/*` → Middleware valida il token contro Neon, 404 se mancante
- `/admin/*` → Auth.js session check, singolo account admin
**Decisioni chiave del data model:**
- `clients.token` è un campo separato e rotazionabile — **non** la primary key
- `clients.accepted_total` denormalizzato: la client API non tocca mai `quote_items`
- `deliverables.approved_at` come audit trail immutabile dal giorno uno
- `payments` sempre due righe per cliente (acconto + saldo), create alla finalizzazione del preventivo
- `ClientView` e `AdminView` sono tipi distinti lato server — privacy enforce a livello di query, non di UI
### Pitfall Critici
1. **Token = primary key (non rotazionabile)** — Usa UUID stabile come PK e campo `token` separato e aggiornabile. Deve essere nella schema della Fase 1; non si può correggere dopo che i link sono stati distribuiti.
2. **Client API espone `quote_items` (nascosti solo nell'UI)** — Definisci `ClientView` come tipo server-side che non interroga mai `quote_items`. Un cliente tecnico con DevTools non deve mai vedere i prezzi singoli.
3. **Over-engineering prima che un cliente usi il portale** — Criterio di successo duro per la Fase 1: un link cliente reale è condivisibile e funziona. Non iniziare il flusso Claude prima che l'admin possa creare un cliente e il cliente possa aprire la sua dashboard.
4. **Nessun record di approvazione immutabile** — Salva `approved_at` (timestamp, non solo boolean) dallo schema iniziale.
5. **Area admin senza vera autenticazione** — Il check Middleware su `ADMIN_PASSWORD` env var deve essere in place prima che la Fase 2 vada in produzione.
---
## Implicazioni per la Roadmap
### Struttura suggerita: 4 fasi
**Fase 1 — Foundation: DB schema, token API, dashboard cliente**
Consegna: un link cliente reale condivisibile che mostra il progetto su mobile e desktop.
Copre: panoramica, fasi/task, pagamenti, documenti, link segreto, DNS.
**Fase 2 — Admin CRUD + auth + commenti + approvazioni**
Consegna: admin crea/modifica clienti, fasi, task, deliverable, pagamenti. Cliente commenta e approva. Admin può rigenerare il link.
Copre: auth Middleware, CRUD completo, flow approvazione con timestamp, commenti lista piatta.
**Fase 3 — Catalogo servizi + preventivi**
Consegna: admin costruisce catalogo riutilizzabile e compone preventivi da esso. `accepted_total` scritto sulla riga cliente.
Nessuna dipendenza client-facing oltre `accepted_total` (già in schema dalla Fase 1).
**Fase 4 (v2) — Flusso Claude AI per onboarding**
Dipende da CRUD stabile + catalogo completo. Claude legge il brief e suggerisce fasi + preventivo.
*Richiede ricerca dedicata durante la pianificazione.*
### Flag di ricerca
- Fasi 13: pattern standard, nessuna ricerca aggiuntiva necessaria
- Fase 4: richiede ricerca su Claude API structured output, streaming vs batch, prompt engineering per generazione fasi
---
## Confidence Assessment
| Area | Confidence | Note |
|------|------------|------|
| Stack | HIGH | Tutte le tecnologie stabili e in produzione |
| Features | HIGH | Feature set opinionated e ben delimitato |
| Architettura | HIGH | Data model completo, pattern two-path auth provato |
| Pitfall | HIGH | Tutti mappabili a decisioni concrete della Fase 1 |
**Domande aperte (da risolvere durante la pianificazione delle fasi):**
- Access log per i link (utile per rilevare accessi non autorizzati)?
- Approvazioni reversibili (admin-only revoke)?
- Log decisioni visibile al cliente dalla v1 o solo admin?
- DNS: configurare e verificare la propagazione nella Fase 1
---
*Ricerca completata: 2026-05-09 | Pronto per la roadmap: sì*