# HelsaGrace Portfolio — Project Casebook

**Document type:** Product design casebook  
**Project:** HelsaGrace Portfolio  
**Owner:** HelsaGrace Okporho  
**Documented:** 16 November 2026  
**Primary evidence:** running application at `http://127.0.0.1:3005`, source repository, case study registry, captured screenshots  
**Companion document:** [PROJECT_BUILD_LOG.md](./PROJECT_BUILD_LOG.md)

> This casebook documents the portfolio as a product: the problem it solves, the structure I chose, the interface decisions behind it and what remains incomplete. It is written from the repository and the running site, not from memory or intention.

---

## 1. Project identity

| Field | Detail |
|---|---|
| **Name** | HelsaGrace Portfolio |
| **Type** | Personal product design and frontend development portfolio |
| **Owner and designer** | HelsaGrace Okporho |
| **Primary statement** | “I design the system. Then I build it.” |
| **Public address** | `https://helsagrace.site` |
| **Repository** | `C:\Users\HelsaGrace\Desktop\My Portfolio\helsagrace` |
| **Framework** | Next.js 15 App Router, React 19, TypeScript strict |
| **Content model** | Two records — Casebook (product design) and Build Log (frontend development) — plus a combined Overview |
| **Current state** | Public portfolio live; content and admin system under active development; new documentation entries added 16 November 2026 |

The portfolio is not a gallery of screenshots. Its job is to show that one person can take a product from an operational problem through interface design to a working build. The site is structured around that claim.

---

## 2. Executive overview

Most design portfolios separate “design” from “development” by showing visual work in one place and a GitHub link in another. That structure makes it hard to see whether the same person carried the decisions through both. This portfolio is built on the opposite premise: **one practice, two records**.

The Casebook contains product design case studies. The Build Log contains frontend builds. The Overview connects them into a continuous view. A single project does not appear in both records, because duplicating a project would blur the distinction the structure is meant to make.

The public site opens with a welcome screen that states the practice and offers three destinations: Casebook, Overview and Build Log. The Casebook is editorial and light. The Build Log is dark and technical. Both share the same brand, typography, navigation and component language, so the split reads as two lenses on one practice rather than two separate websites.

**Status in one line:** The public portfolio is live and the two-record structure is implemented; the content system continues to evolve, and some project documents describe older states of the site.

---

## 3. Context

### 3.1 The professional context

I work across product design and frontend development. The roles are often hired separately, and portfolios usually force a choice: show polished interfaces and let the code be inferred, or show code and let the design be inferred. Neither shows the handoff, which is where most product intent is lost.

The portfolio has to make a case for the combination without claiming that the two disciplines are the same. The structure therefore separates the records but keeps them in one system.

### 3.2 The content context

The portfolio contains work across fintech, developer tools, skincare, career technology, live music, local commerce, food delivery and lending. The work ranges from full product builds to concept case studies. A single template would flatten that range.

The two-record model lets each project use the lens that fits it. A concept case study belongs in the Casebook. A shipped product belongs in the Build Log. A project is not forced into both.

### 3.3 The technical context

The public site is statically rendered from a content snapshot and a case study registry. Supabase provides the admin content store, authentication and row-level security. The site is designed to work publicly without Supabase configuration; the admin requires it.

### 3.4 The documentation context

At documentation time the repository's `CONCEPT.md` describes some routes that have since changed (`/explore` now redirects to `/`), and the README names a body typeface that the current layout no longer uses. The code is authoritative. Those discrepancies are recorded in section 19 rather than repeated as fact.

---

## 4. The problem

### 4.1 The perception problem

A portfolio that shows only finished screens invites the assumption that someone else built them. A portfolio that shows only code invites the assumption that the interface is incidental. I needed a structure where the design decisions and the production decisions are both visible, and where the connection between them is explicit.

### 4.2 The navigation problem

A visitor arrives with a role and a limited amount of time. A design lead wants to see how I frame a problem. A frontend lead wants to see how I structure a codebase. A founder wants to know whether I can do both. The site cannot assume one reading order.

### 4.3 The evidence problem

Case studies often claim outcomes that cannot be verified. I wanted the portfolio to distinguish between a shipped product, a concept, and a build in progress. The status vocabulary — Live, In build, Prototype, Concept — exists for that reason, and it is applied honestly to each entry.

### 4.4 The range problem

The work spans full products and focused concepts. A single long-form template would overstate the concepts and understate the builds. The two-record structure preserves the difference in depth.

### 4.5 What I did not try to solve

I did not build a public blog, a client portal, a CMS marketplace or a social feed. The portfolio is a focused argument, not a platform.

---

## 5. Goals and non-goals

### 5.1 Goals

| Goal | How the site expresses it |
|---|---|
| Show design and build as one practice | Two records, one overview, shared brand and components |
| Let visitors choose the lens they need | Welcome screen with three explicit destinations |
| Preserve the difference between records | Casebook is light and editorial; Build Log is dark and technical |
| Make each project’s status legible | Explicit status labels and honest copy |
| Support depth without a wall of text | A block model with narrative, steps, galleries, systems and reflections |
| Keep the public site fast and resilient | Static rendering from local content; no backend required to read |
| Keep the admin private | Supabase auth, row-level security and an owner allowlist |
| Make new entries additive | Registry-driven routes, sitemap and counts |

### 5.2 Non-goals

- A single grid of thumbnails with no narrative.
- Duplicating the same project in both records.
- A visual theme per project that breaks the shared brand.
- Anonymous public comments or social features.
- A public admin sign-up.
- Publishing private product surfaces (`/os`, `/devotional`, `/recovery`, `/vela`).

The non-goals protect the structure. The split only works if the records stay distinct and the brand stays coherent.

---

## 6. Users and roles

### 6.1 The design hiring lead

They want to understand how I frame a problem, structure information and make interface decisions. They read the Casebook first. Their critical question is whether the thinking is specific or decorative.

### 6.2 The engineering hiring lead

They want to see architecture, data modelling, state, accessibility, testing and release discipline. They read the Build Log first. Their critical question is whether the build is real.

### 6.3 The founder or product lead

They want evidence that one person can carry a product from problem to production. They may read the Overview, which connects both records.

### 6.4 The peer or collaborator

They want to see how the work is made and what I have learned. They may move between records and read the reflections.

### 6.5 The owner (me)

I need a system I can extend without redesigning it for every new project. The registry, block model, archive and documentation sync exist to make additions additive.

### 6.6 Access model

| Role | Surface | Gate |
|---|---|---|
| Visitor | Welcome, Casebook, Build Log, Overview, study pages | None |
| Owner | `/admin/*` | Supabase session plus owner allowlist |
| Private product user | `/os`, `/devotional`, `/recovery`, `/vela` | Owner or product-specific access; outside public portfolio scope |

The private products are deliberately excluded from the public case study system. They are mentioned only as separate surfaces, not documented in detail.

---

## 7. Product strategy

### 7.1 Structure as the argument

The site's main idea is its structure. Instead of claiming “I do both”, it gives each discipline a record and then shows the connection in the Overview. The structure does the persuading.

### 7.2 One brand, two temperatures

Both records share the same header, footer, typography, purple accent and component vocabulary. The Casebook uses the off-white editorial surface; the Build Log uses a dark graphite surface. The difference is a change of temperature, not a change of identity.

### 7.3 No duplicate projects

The same project does not appear in both records. This is a deliberate constraint. Duplication would suggest that the two records are lenses on identical content, when they are actually different kinds of evidence.

### 7.4 Depth by block, not by page length

Case studies are composed from blocks: narrative, steps, quote, gallery, architecture, reflection, next and system. The block model lets a study carry the right amount of detail without a bespoke page layout for every project.

### 7.5 Additive growth

New studies are appended to the registry. Routes, sitemap and counts derive from that registry. Adding a project should not require a new template or a navigation redesign.

### 7.6 Documentation as part of the product

The portfolio now carries project documentation as first-class content: casebooks, build logs and portfolio entries generated from Markdown. The documentation is part of the evidence, not an afterthought.

---

## 8. Feature set

| Area | Feature | Status |
|---|---|---|
| Entry | Welcome screen with three destinations | Implemented |
| Casebook | Hub with featured cases and counts | Implemented |
| Casebook | Individual case study pages | Implemented |
| Casebook | Archive | Implemented |
| Build Log | Hub with featured builds and counts | Implemented |
| Build Log | Individual build pages | Implemented |
| Build Log | Archive | Implemented |
| Overview | Combined practice view | Implemented |
| Study system | Block model: narrative, steps, quote, metrics, gallery, architecture, reflection, next, system | Implemented |
| Study system | Contents rail, next-study navigation, cross-record link | Implemented |
| Study system | Optional product accent inside a study | Implemented |
| Study system | Responsive image fitting and captions | Implemented |
| Content | Content snapshot with typed merge | Implemented |
| Content | Supabase admin content store and revisions | Implemented; live admin not exercised in this documentation pass |
| Content | Documentation sync from Markdown to TypeScript | Implemented 16 November 2026 |
| Discovery | Sitemap generated from the registries | Implemented |
| Discovery | Legacy `/work/[slug]` pages | Implemented |
| Contact | Contact form with Resend, honeypot and rate limiting | Implemented; requires configuration |
| Accessibility | Skip link, focus management, reduced motion | Implemented |
| Private products | `/os`, `/devotional`, `/recovery`, `/vela` | Implemented separately; excluded from public case studies |

---

## 9. Information architecture

### 9.1 Public architecture

```
/                         Welcome: the practice and three destinations
/casebook                 Product design record
/casebook/archive         All case studies
/casebook/[slug]          A case study
/build-log                Frontend development record
/build-log/archive        All builds
/build-log/[slug]         A build study
/overview                 Both records in one continuous view
/work/[slug]              Legacy project pages
/resume                   Résumé
/resume/[role]            Role-specific résumé
```

### 9.2 Private architecture (outside public case studies)

```
/admin/*                  Owner content administration
/os/*                     Personal operating system
/devotional/*             Devotional product
/recovery/*               Recovery product
/vela/*                   Messaging product
```

These surfaces are mentioned for completeness. They are not part of the portfolio's public case study scope, and no private content is reproduced in this documentation.

### 9.3 Record architecture

Each record is a hub plus an archive plus study pages. The hub shows the featured set and the total count. The archive holds everything. Study pages are generated from the registry by slug.

This gives each record a clear front door without turning the hub into a long list.

### 9.4 Navigation

The header offers Overview, Casebook and Build Log, plus Contact. The mobile menu uses the same destinations with focus trapping and Escape handling. The footer repeats the record links and adds legal and contact routes.

The welcome screen is intentionally outside that header pattern: it is an entry cover, not a record page.

---

## 10. Key user flows

### 10.1 Choose a lens

1. Visitor lands on `/`.
2. They read the practice statement and the short explanation of the two records.
3. They choose Casebook, Overview or Build Log.
4. The chosen hub shows featured work and the total count.

### 10.2 Read a design case

1. Visitor opens a Casebook study.
2. They see the cover, title, subtitle, role, category, status and year.
3. A contents rail lists the study sections.
4. They scroll through narrative, steps, galleries and the interface system.
5. They reach a reflection and a next link.
6. They can move to the next case or return to the record hub.

### 10.3 Read a build

1. Visitor opens a Build Log study.
2. They see the dark technical surface, cover and metadata.
3. The contents rail exposes the brief, architecture, system map, hard parts, proof, screens and reflection.
4. They read the engineering decisions and see the production screens.
5. They can cross to the other record or open the live product where one exists.

### 10.4 Read both

1. Visitor opens `/overview`.
2. They see the combined practice statement, counts and both sets of work.
3. They can enter either record from the same page.

### 10.5 Extend the portfolio (owner)

1. Add a study to the registry or to a project's `PORTFOLIO_CONTENT.md`.
2. Run the documentation sync.
3. The route, sitemap entry and counts update from the registry.
4. Verify the study in the browser and the production build.

---

## 11. UX and interface decisions

### 11.1 Three destinations, not a toggle

The earlier concept used a lens toggle on the same project. The current model removes it. A toggle implies that the same content can be viewed two ways; the records actually contain different projects and different kinds of evidence. Three explicit destinations are clearer and more honest.

### 11.2 The cover belongs to both records

The welcome screen is one cover, not a dark/light split. It establishes the practice and offers the branches. The visual divergence begins after the choice, which makes the choice feel like entering a record rather than switching a theme.

### 11.3 Editorial Casebook, technical Build Log

The Casebook uses the same off-white base as the rest of the portfolio, generous whitespace and a serif-led hierarchy. The Build Log uses a dark graphite surface, browser frames and architecture diagrams. The contrast signals the change in evidence without changing the brand.

### 11.4 Status is explicit

Each study carries a status: Live, In build, Prototype or Concept. The status is not decorative. It tells the reader how much of the work is shipped and how much is proposed.

### 11.5 Contents rails reduce long-page disorientation

Long studies expose a contents rail with section labels. The reader can see the shape of the argument before committing to it, and can move between sections without losing context.

### 11.6 Images are evidence, not decoration

Study images carry alt text and captions. Landscape and portrait images use different aspect ratios. Galleries can switch between grid and marquee depending on the set. The intent is that every image explains something.

### 11.7 Metrics are used sparingly

The block model supports metrics, but the study view filters metrics blocks out of rendered content in the current implementation. Metrics still appear in the hero and footer of the broader portfolio, where they are part of the site's summary rather than a case study section. New studies should not rely on a metrics block to communicate essential information.

### 11.8 Motion is restrained

GSAP and Lenis provide reveal and scroll behaviour, but the interface respects reduced motion and never depends on animation to convey state. The design should read with motion disabled.

### 11.9 Accessibility is structural

Skip links, landmarks, visible focus, labelled controls, keyboard-operable menus and focus restoration are part of the shell. The portfolio teaches accessibility in its case studies; the site has to meet the same standard.

---

## 12. Visual direction and design system

### 12.1 Direction

The visual language is editorial and typographic. Large serif headlines carry the argument; a humanist sans carries the reading; a mono face marks labels, metadata and technical detail. The palette is a warm off-white and ink with a single purple accent, plus a dark graphite surface for the Build Log.

The result is closer to a designed publication than a template portfolio. That is intentional: the site is making an argument, and the typography is part of the argument.

### 12.2 Palette

| Token | Value | Role |
|---|---|---|
| `--path-bg` | `#f9f9fb` | Light record background |
| `--path-surface` | `#ffffff` | Cards and panels |
| `--path-ink` | `#0f0f14` | Primary text |
| `--path-muted` | `#6b6b7a` | Secondary text |
| `--path-border` | `#e4e4e9` | Dividers and card edges |
| `--path-accent` | `#5b4fe8` | Brand purple |
| `--path-accent-soft` | `#eeecff` | Soft accent surface |
| `--path-shot-bg` | `#f0f0f5` | Image backing |
| Dark `--path-bg` | `#0f0f14` | Build Log background |
| Dark `--path-surface` | `#17171f` | Build Log panels |
| Dark `--path-ink` | `#f9f9fb` | Text on dark |
| Dark `--path-muted` | `#a1a1ad` | Secondary text on dark |
| Dark `--path-accent` | `#a79fff` | Lavender accent on dark |

Defined in `app/concept.css`.

### 12.3 Typography

| Role | Family | Usage |
|---|---|---|
| Display | Instrument Serif | Headlines and study titles |
| Body | Manjari | Reading copy and interface text |
| Mono | IBM Plex Mono | Labels, metadata, section indices |

The display face uses tight tracking and italic emphasis for the second half of key statements (“I design the system. *Then I build it.*”). Mono labels are uppercase with wide tracking. Body copy is set at a comfortable measure with a generous line height.

### 12.4 Components

- Shared header and footer across both records.
- Record hub with a large statement, counts and featured cards.
- Study card with cover, kicker, title, subtitle, metadata and a read link.
- Contents rail with numbered section labels.
- Narrative, steps, quote, gallery, architecture, system, reflection and next blocks.
- Image component with aspect ratio, fit, caption and optional gallery overlay.
- Mobile menu with focus trap and Escape handling.

### 12.5 The two temperatures

| | Casebook | Build Log |
|---|---|---|
| Surface | Off-white | Dark graphite |
| Tone | Editorial, reflective | Technical, evidentiary |
| Imagery | Product screens and interface detail | Browser frames, architecture, production screens |
| Accent | Purple on light | Lavender on dark |
| Status | Concept, Prototype, Live | Live, In build |

### 12.6 What the system avoids

- A separate visual brand for each record.
- Project-specific themes that break the shell.
- Decorative gradients without a structural role.
- Animation that hides content or delays reading.
- Metrics that cannot be supported by evidence.

---

## 13. Screen inventory

All screenshots were captured on 16 November 2026 from the local application at `http://127.0.0.1:3005`, served from a production build (`next build` + `next start`), using isolated headless Chrome with no authenticated session. Desktop viewport: 1440 × 1000. Mobile viewport: 390 × 844. Every captured page reported zero horizontal overflow and zero runtime errors. The base manifest is [`PROJECT_ASSETS/supporting-evidence/capture-manifest.json`](./PROJECT_ASSETS/supporting-evidence/capture-manifest.json); the integration run is [`PROJECT_ASSETS/supporting-evidence/integration-browser-checks.json`](./PROJECT_ASSETS/supporting-evidence/integration-browser-checks.json).

### 13.1 Desktop

| Screen | Image | What it shows |
|---|---|---|
| Welcome | [`screenshots/home-desktop.png`](./PROJECT_ASSETS/screenshots/home-desktop.png) | The practice statement and the three destination cards |
| Casebook hub | [`screenshots/casebook-desktop.png`](./PROJECT_ASSETS/screenshots/casebook-desktop.png) | The Product Design statement, counts and the first case card |
| Build Log hub | [`screenshots/build-log-desktop.png`](./PROJECT_ASSETS/screenshots/build-log-desktop.png) | The dark Frontend Development statement and the first build card |
| Overview | [`screenshots/overview-desktop.png`](./PROJECT_ASSETS/screenshots/overview-desktop.png) | The combined practice, counts and the relationship between the records |
| Build study | [`screenshots/careercraft-desktop.png`](./PROJECT_ASSETS/screenshots/careercraft-desktop.png) | A live Build Log study with metadata, contents rail and the opening brief |

### 13.2 Mobile

| Screen | Image | What it shows |
|---|---|---|
| Welcome | [`mobile/home-mobile.png`](./PROJECT_ASSETS/mobile/home-mobile.png) | The practice statement and stacked destination cards |
| Casebook hub | [`mobile/casebook-mobile.png`](./PROJECT_ASSETS/mobile/casebook-mobile.png) | The Casebook statement, counts and first card at 390px |
| Build Log hub | [`mobile/build-log-mobile.png`](./PROJECT_ASSETS/mobile/build-log-mobile.png) | The dark Build Log hub at 390px |
| Overview | [`mobile/overview-mobile.png`](./PROJECT_ASSETS/mobile/overview-mobile.png) | The combined view at 390px |
| Build study | [`mobile/careercraft-mobile.png`](./PROJECT_ASSETS/mobile/careercraft-mobile.png) | A Build Log study at 390px |

### 13.3 Integration evidence

A second capture pass recorded the four documented studies, the updated hubs and both archives. It also re-captured two existing studies — CareerCraft and OPay — to show that integration did not alter them.

| Screen | Image | What it shows |
|---|---|---|
| Casebook hub | [`screenshots/casebook-hub-updated-desktop.png`](./PROJECT_ASSETS/screenshots/casebook-hub-updated-desktop.png) | Counts now read 05 product design cases and 07 frontend builds |
| Build Log hub | [`screenshots/build-log-hub-updated-desktop.png`](./PROJECT_ASSETS/screenshots/build-log-hub-updated-desktop.png) | The dark hub with the expanded build library |
| Casebook archive | [`screenshots/casebook-archive-updated-desktop.png`](./PROJECT_ASSETS/screenshots/casebook-archive-updated-desktop.png) | Five design entries |
| Build Log archive | [`screenshots/build-log-archive-updated-desktop.png`](./PROJECT_ASSETS/screenshots/build-log-archive-updated-desktop.png) | Seven build entries |
| CraftCode casebook study | [`screenshots/academy-casebook-desktop.png`](./PROJECT_ASSETS/screenshots/academy-casebook-desktop.png) | The new design study with metadata and contents rail |
| CraftCode build log study | [`screenshots/academy-build-log-desktop.png`](./PROJECT_ASSETS/screenshots/academy-build-log-desktop.png) | The paired engineering study |
| Portfolio casebook study | [`screenshots/portfolio-casebook-desktop.png`](./PROJECT_ASSETS/screenshots/portfolio-casebook-desktop.png) | The study describing the two-record structure |
| Portfolio build log study | [`screenshots/portfolio-build-log-desktop.png`](./PROJECT_ASSETS/screenshots/portfolio-build-log-desktop.png) | The study describing this site's implementation |
| CareerCraft study | [`screenshots/careercraft-preserved-desktop.png`](./PROJECT_ASSETS/screenshots/careercraft-preserved-desktop.png) | Existing study unchanged |
| OPay case | [`screenshots/opay-preserved-desktop.png`](./PROJECT_ASSETS/screenshots/opay-preserved-desktop.png) | Existing case unchanged |

The same ten states were captured at 390 px in the `mobile/` directory with matching `-mobile.png` filenames.

### 13.4 Screens not captured, and why

| Screen | Reason |
|---|---|
| `/admin/*` | Requires an owner session. No credentials were used. |
| `/os`, `/devotional`, `/recovery`, `/vela` | Private product surfaces; outside public portfolio scope. |
| Contact success state | Requires a configured Resend key and a real submission. |
| Admin editor states | Require an owner session and would expose private drafts. |

---

## 14. Technical architecture, from a design perspective

The portfolio is a Next.js App Router site. Its design-relevant properties are:

- **Static public reading.** The public pages render from a local content snapshot and a typed case study registry. A visitor does not need Supabase to read the portfolio.
- **A shared block renderer.** Every study page is composed from the same blocks, so a new project can be added without a new page layout.
- **Separate record shells.** The route group applies the concept stylesheet and the shared header/footer. The Build Log's dark temperature is a surface decision inside that shell.
- **Registry-driven discovery.** Hubs, archives, counts, next-study links and the sitemap derive from the same registries. Adding a study updates them automatically.
- **Markdown-authoritative documentation.** Project documentation lives in Markdown; a sync script generates typed entries so the site and the documents cannot drift silently.
- **Private administration.** Supabase auth and row-level security gate the admin. The public site does not depend on that connection.

A full engineering account is in the [build log](./PROJECT_BUILD_LOG.md).

---

## 15. Key decisions and trade-offs

### 15.1 Two records instead of one mixed grid

**Decision:** Separate Casebook and Build Log, with an Overview that connects them.

**Why:** The two kinds of evidence are different. Mixing them forces every project into one template and makes it harder for a specialist reader to find what they need.

**Trade-off:** More navigation and more content discipline. The welcome screen has to explain the structure in a few lines.

### 15.2 No duplicate projects

**Decision:** A project belongs to one record.

**Why:** Duplication would undermine the distinction and inflate the portfolio's apparent size.

**Trade-off:** Some projects could plausibly sit in either record. The choice has to be made deliberately.

### 15.3 Shared brand, different temperatures

**Decision:** One typographic and colour system; two surface treatments.

**Why:** The records should feel like two lenses on one practice, not two brands.

**Trade-off:** The Build Log cannot use a completely different visual language, even where a project might benefit from it. Product accents are allowed inside a study, but the shell stays consistent.

### 15.4 Block model instead of bespoke pages

**Decision:** Compose studies from a fixed block set.

**Why:** It makes new work additive and keeps the reading experience consistent.

**Trade-off:** A project with an unusual story may need a block combination that does not fit neatly. The model is broad enough for the current work but not infinitely flexible.

### 15.5 Static content with an optional CMS

**Decision:** Public pages read from a local snapshot; the admin writes to Supabase.

**Why:** The portfolio stays fast and available even when the backend is not configured.

**Trade-off:** Content can exist in two places during development, and the sync between them has to be managed.

### 15.6 Documentation in Markdown, generated into the app

**Decision:** Keep the authoritative project documentation in Markdown and generate the portfolio entries from it.

**Why:** The user asked for Markdown to remain authoritative. Generating the entries prevents the site and the documents from diverging.

**Trade-off:** An extra build step and a validation check. Editing the site data by hand is no longer the source of truth.

### 15.7 Exclude private products from public case studies

**Decision:** Mention `/os`, `/devotional`, `/recovery` and `/vela` as separate surfaces but do not document their contents publicly.

**Why:** They contain private or owner-specific material. The public portfolio does not need them to make its case.

**Trade-off:** Some of the most interesting engineering is not shown in detail. The boundary is deliberate.

---

## 16. Challenges and constraints

### 16.1 Two audiences, one site

A design lead and an engineering lead read differently. The structure has to serve both without making either feel like they are in the wrong place. The welcome screen and the Overview are the compromise: one entry, three reading paths.

### 16.2 Content range

The work includes shipped products and concepts. The status vocabulary and the two records preserve the difference, but the range still puts pressure on the block model.

### 16.3 Documentation drift

`CONCEPT.md` and the README describe older route and typeface choices. The code is current. Documentation drift is a real maintenance cost, and it is one reason the new project documentation is generated rather than hand-copied.

### 16.4 Production screenshots

Case studies depend on real product screens. Some source images are large, some are portrait, and some were captured at different times. The image component handles aspect ratio and fit, but the content still needs curation.

### 16.5 Honest metrics

The site's counts and metrics have to come from the registry or from real evidence. A portfolio that inflates its numbers is easy to disprove. The counts on the hubs are derived from the actual arrays.

### 16.6 Private content boundaries

The portfolio sits alongside private products in the same repository. The public case study system has to stay clearly separated from those surfaces, in routes, sitemap and documentation.

### 16.7 No research archive

There is no folder of user interviews or usability findings for the portfolio itself. The design rationale documented here comes from the shipped structure, the copy and the implementation. I do not claim formal research that was not recorded.

---

## 17. Testing and evidence

### 17.1 Automated tests

The repository contains five test files:

- `tests/time/wat.test.ts` — timezone handling
- `tests/email/weekly-review.test.ts` — weekly review email logic
- `tests/os/profile.test.ts` — OS profile logic
- `tests/os/review.test.ts` — OS review logic
- `tests/os/tasks.test.ts` — OS task logic

These cover shared logic and private product behaviour. The portfolio's case study registry and documentation sync are verified by type checking, the sync check and browser inspection rather than by unit tests.

### 17.2 Build and type checks

The project uses TypeScript strict mode, ESLint and a production build. The build log records the validation run for this documentation session.

### 17.3 Browser evidence

Thirty screenshots were captured from the production build: five desktop and five mobile for the base set, plus ten desktop and ten mobile for the integration set. The harness waited for page readiness and verified:

- fonts loaded;
- images loaded;
- finite animations completed;
- two animation frames rendered;
- no runtime errors;
- no horizontal overflow at either viewport.

The base set covers the welcome screen, both hubs, the Overview and one Build Log study. The integration set covers both hubs, both archives, all four documented studies and two preserved studies.

### 17.4 Navigation checks

The capture set includes desktop and mobile views of the hubs and a study page. The mobile header exposes the menu trigger and the Contact action. The study page shows the back link, status pill, metadata and contents rail, confirming the study shell renders at both widths.

### 17.5 What the evidence does not cover

- No authenticated admin session was exercised.
- No contact submission was sent.
- No formal accessibility audit (axe or screen reader) was recorded.
- No performance or load test was run.
- No analytics or user behaviour data exists in the repository.
- The archive pages were captured, but their filters were not exercised.

---

## 18. Current status

| Surface | Status | Evidence |
|---|---|---|
| Welcome screen | Live | Public site, captured desktop and mobile |
| Casebook hub and studies | Live | Public site, registry, captured hub |
| Build Log hub and studies | Live | Public site, registry, captured hub and study |
| Overview | Live | Public site, captured desktop and mobile |
| Archives | Live | Source routes; captured at both widths in the integration set |
| Sitemap | Implemented | Generated from both registries |
| Admin content system | Implemented; requires Supabase | Source; not exercised in this pass |
| Contact | Implemented; requires Resend | Source; not exercised in this pass |
| Documentation sync | Implemented 16 November 2026 | Script and generated registry |
| Four new documentation entries | Added 16 November 2026 | `PORTFOLIO_CONTENT.md` files and generated data |

**Content counts before this documentation pass:** three Casebook studies and five Build Log studies. After the pass: five and seven. The counts are derived from the registries, not typed by hand.

---

## 19. Limitations and known defects

### 19.1 Documentation drift

1. **`CONCEPT.md` route notes are stale.** It describes `/` as a welcome screen with a CTA to `/explore`, but `/explore` now redirects to `/` and the welcome screen offers its three destinations directly.
2. **The README and CONCEPT name different body typefaces.** The README says Instrument Serif + Inter; CONCEPT says Instrument Serif + Inter + IBM Plex Mono. The current layout imports Instrument Serif, IBM Plex Mono and Manjari, and uses Manjari as the primary sans. The code is authoritative.
3. **The CONCEPT template description includes metrics in the study flow.** The current study view filters metrics blocks out of rendered content. New studies should not depend on them.

### 19.2 Product limitations

- The admin content system is implemented but was not exercised in this documentation pass because it requires owner credentials.
- The contact form requires Resend configuration; no live submission was tested.
- The archive pages exist but were not included in the capture set.
- The site has no search or filtering across the records.
- The site has no public RSS or writing section.
- Private products are excluded from the public case study system, so the portfolio does not show that part of the engineering work in detail.

### 19.3 Evidence limitations

- No user research archive exists for the portfolio.
- No analytics data, conversion data or interview findings exist in the repository.
- No formal accessibility audit or performance budget is recorded.
- The captured study page uses CareerCraft's cover image, which rendered as a very light frame in the headless capture; the page structure is visible, but the image content is not a reliable representation of the product screen.

### 19.4 Content risks

- Case study claims must remain supportable. Metrics that cannot be traced to evidence should not be added.
- New entries should preserve the honesty of the status vocabulary. “Live” should mean deployed and usable, not “built locally”.
- The four new documentation entries describe two projects that are still evolving; their status must be updated when the underlying products change.

---

## 20. Future roadmap

### 20.1 Content

1. Add remaining project case studies using the block model.
2. Keep the status of each study current as products ship.
3. Replace any unsupported metric with a traceable one or remove it.
4. Add a short writing or notes section if the practice needs one.

### 20.2 Product

1. Add filtering or search to the archives as the library grows.
2. Make the contents rail sticky and keyboard-navigable on long studies.
3. Add previous/next study navigation within a record.
4. Improve the study image pipeline for large and portrait source screens.
5. Add a visible documentation link convention across all studies.

### 20.3 Administration

1. Exercise the Supabase admin end to end with an owner session.
2. Confirm content revisions and media handling in production.
3. Decide whether the content snapshot or the database is the long-term source of truth for public pages.

### 20.4 Quality

1. Add tests for the case study registry and documentation sync.
2. Run an accessibility audit and fix findings.
3. Add a performance budget for the study pages.
4. Capture the archive pages in the evidence set.

---

## 21. Reflection

The portfolio's central design decision was to stop treating design and development as two portfolios joined by a link. The two-record structure costs more navigation and more content discipline, but it makes the practice legible in a way a mixed grid cannot.

The decision I am least settled on is the depth of the Build Log. Engineering evidence is valuable, but a build log can become a second résumé. The block model helps by forcing each study to choose a few arguments rather than listing everything.

The part I am most confident about is the shared brand. The Casebook and Build Log feel different without feeling like different people made them. That is the point: one practice, two lenses.

The next improvement is not visual. It is evidentiary. The portfolio should be able to show, for every “Live” claim, where the product is live and what changed. That is the same standard I applied to the project documentation, and it should apply to the portfolio itself.

---

## 22. Summary

The HelsaGrace Portfolio is a personal product design and frontend development portfolio built around one practice and two records. The Casebook carries design case studies; the Build Log carries frontend builds; the Overview connects them. The public site is live, static and fast, with a shared brand and a block-based study system that makes new work additive.

The portfolio's strongest asset is its structure. It does not ask a visitor to believe that design and development belong together; it shows the two kinds of evidence side by side and lets the connection speak. Its main weakness is documentation drift and the absence of formal research or analytics evidence, both of which are recorded here rather than hidden.

The documentation added on 16 November 2026 extends the portfolio with two casebooks and two build logs, generated from Markdown so that the published entries and the authoritative documents stay aligned.

---

### Evidence index

| Evidence | Location |
|---|---|
| Case study registry | `lib/case-studies.ts` |
| Portfolio content fixture | `lib/concept-content.ts`, `content-snapshot.json` |
| Study renderer | `components/concept/case-study-view.tsx` |
| Record hubs | `components/concept/path-hub.tsx`, `components/concept/work-section.tsx` |
| Welcome screen | `components/welcome/welcome-screen.tsx` |
| Design tokens | `app/concept.css` |
| Routes | `app/(welcome)`, `app/(records)`, `app/overview` |
| Sitemap | `app/sitemap.ts` |
| Documentation sync | `scripts/sync-project-documentation.mjs` |
| Screenshots and manifest | [`PROJECT_ASSETS`](./PROJECT_ASSETS) |
| Engineering record | [PROJECT_BUILD_LOG.md](./PROJECT_BUILD_LOG.md) |
