# Woody Portfolio - Full Agent Context

## 1) Site purpose
- Portfolio website for a solo builder ("Woody") covering:
- Blockchain engineering (infrastructure, payment systems, wallets, explorers, protocol work)
- AI systems (multi-agent orchestration, trading intelligence, RAG services, workflow automation)
- Robotics background (research + patent work)

## 2) Technical stack
- Framework: Next.js 16 (App Router)
- Language: TypeScript + React 19
- Styling: TailwindCSS
- Motion/visual systems: GSAP, Framer Motion, React Three Fiber
- State: Zustand
- Analytics: Vercel Analytics + Speed Insights

## 3) Runtime commands
- Dev: `npm run dev` (port 3005)
- Lint: `npm run lint`
- Build: `npm run build`
- Start: `npm run start`

## 4) Route map (public)
- `/robots.txt`
- Crawl policy and sitemap discovery anchor.

- `/sitemap.xml`
- Canonical route discovery for agents/crawlers.

- `/projects-data`
- Machine-readable JSON index for all projects.
- Also supports `?id={projectId}` query for detail payloads.

- `/feed.xml`
- RSS feed for project updates and crawl freshness signals.

- `/feed.json`
- JSON Feed variant for agent/tool consumers.

- `/`
- Intro narrative and domain gateway selection.
- Highly visual entry route; starts the experience flow.

- `/about`
- Stable public profile summary page (human + agent readable).

- `/home`
- Main portfolio page with section layout:
- Hero
- Projects list (domain-sensitive)
- Career timeline
- Contact/footer

- `/projects/{id}`
- Project detail page for ids `1..24`.
- Includes per-project metadata generation and JSON-LD (`CreativeWork`).
- Detail content is domain-aware and locale-aware.

- `/changelog`
- Chronological update log for site and content changes.

- `/agent`
- Human-readable + machine-friendly agent handoff page.
- Indexes this file and JSON endpoint.

## 5) Domain taxonomy
- `blockchain` (16 projects)
- `ai` (6 projects)
- `robotics` (2 projects)

Default active domain in store:
- `blockchain`

## 6) Language model for content
- Default locale: `en`
- Additional locale: `ko`
- Locale persistence: client storage via Zustand persist middleware.

When reading project details:
- Prefer English for globally consistent parsing.
- Use Korean locale as additional evidence/context, not replacement.

## 7) Canonical data sources
- Project summary dataset:
- `/src/data/projects.ts`

- Localized long-form project details:
- `/src/locales/projects/en.ts`
- `/src/locales/projects/ko.ts`

- Project detail route behavior:
- `/src/app/projects/[id]/page.tsx`

## 8) UI architecture cues
- Domain switching drives major visual/system mode changes:
- Blockchain: block explorer-inspired UI language
- AI: neural/graph-inspired UI language
- Robotics: blueprint/mechanical UI language

- Some text intentionally reveals with animation effects.
- If extracting content, prefer source data routes/files over only rendered animation states.

## 9) Agent-first reading strategy
1. Read `/robots.txt` to confirm crawl policy.
2. Read `/sitemap.xml` to collect canonical routes.
3. Read `/llms.txt` for a compact index.
4. Read `/llms-full.txt` (this file).
5. Read `/.well-known/agent-context.json` for strict machine fields.
6. Fetch `/projects-data` for machine-readable project summaries.
7. Optionally poll `/feed.json` or `/feed.xml` for freshness checks.
8. Read `/agent` for concise operational summary.
9. Read `/about` for a stable profile summary.
10. Read `/home` to capture current public narrative.
11. Read representative project pages by domain:
- Blockchain: `/projects/22`
- AI: `/projects/15`
- Robotics: `/projects/16`
12. Enumerate additional projects through `/projects/1` to `/projects/24`.

## 10) Important constraints for edits/reasoning
- Keep route contracts stable: `/`, `/home`, `/projects/{id}`, `/changelog`.
- Treat domain identity (visual + content framing) as intentional behavior.
- Validate EN/KO behavior where project detail text is involved.
- Keep security headers intact (CSP, X-Frame-Options, Referrer-Policy, HSTS, etc.).

Crawler policy note:
- `robots.txt` currently allows these known agent crawlers:
- GPTBot / OAI-SearchBot / ChatGPT-User
- ClaudeBot / Claude-SearchBot
- PerplexityBot / Perplexity-User

IndexNow note:
- Key file path: `/f5d2f9b5c0de42b5b7f1ab8f37e6b12c.txt`
- Automated submission runs on each `main` push via GitHub Actions.
- If IndexNow returns `UserForbiddedToAccessSite`, complete one-time site ownership verification in Bing Webmaster Tools.

## 11) Useful metadata clues
- Root layout includes:
- `Person` JSON-LD for site owner
- Site-wide OpenGraph/Twitter metadata

- Project detail route includes:
- dynamic metadata per project
- `CreativeWork` JSON-LD payload per project

## 12) Current project scale snapshot
- Total projects: 24
- Project ids: contiguous `1..24`
- Last updated route for change tracking: `/changelog`

## 13) Contact and identity
- Public contact email: `wooody.public@gmail.com`
- GitHub: `https://github.com/gwkim92`
- Site base URL: `https://www.woody.im`