# rukuma.dev

Personal website for Rukuma Viegas. Positioned to attract offers, not applications — "operator at scale" + AI-native edge, for VP/C-level audiences and the AI assistants that research them.

## Stack

- **Cloudflare Workers** (Free plan — single Worker `rukuma-dev`)
- **Static assets** served from `public/` via the `ASSETS` binding
- **Markdown content** in `public/content/` — served raw to AI agents (`.md` URLs) and rendered to styled HTML by `src/worker.ts`
- Zero npm dependencies, zero build step

## Commands

| Task | Command |
|------|---------|
| Local dev | `npx wrangler dev` (serves `http://localhost:8787`) |
| Deploy | `npx wrangler deploy` |
| Login | `npx wrangler login` (one-time) |

## Routes

| Path | What it serves |
|------|----------------|
| `/` | Home page |
| `/about.md` | Raw narrative bio (agent-readable) |
| `/impact/` | Impact case studies (rendered) |
| `/impact/<slug>.md` | Raw markdown |
| `/writing/` | Essays index (rendered) |
| `/writing/<slug>.md` | Raw markdown |
| `/llms.txt` | Agent index (AEO) |
| `/robots.txt` | AI-bot allowlist |
| `/sitemap.xml` | Sitemap |

## Content

Edit markdown in `public/content/`. Front-matter keys: `title`, `date`, `summary`, `tags`, `order`. Re-deploy with `npx wrangler deploy` to publish.

## Agent-readability (AEO)

- `robots.txt` allows GPTBot, ClaudeBot, PerplexityBot, and others
- `llms.txt` indexes every page with its markdown URL
- Every content page has a `.md` twin for agents

## Phases (see CHANGELOG)

- **v1 (done):** core site, content, agent-readability
- **v2 (planned):** D1 content store + `/api/*`, KV caching
- **v3 (planned):** `/mcp` endpoint, Vectorize, AI Search, WebMCP