d7ed55399893be3727b0416a67f933178b8c2e29
new file mode 100644
index 0000000..455753f
@@ -0,0 +1,230 @@
+---
+tags:
+- agi-house
+- llm-wiki
+- karpathy
+- farzapedia
+- event
+title: AGI House — LLM Wiki Event
+visibility: public-edit
+---
+
+# AGI House — LLM Wiki Event
+
+> Public-editable collaborative wiki for the AGI House **LLM Wiki** event.
+> **Anyone can edit any page here** — no account needed. Add notes, links,
+> tips, your own wiki, whatever's useful.
+
+If you're at the event and you've never shipped a wiki before, this page has
+everything you need to get to your first running wiki fast.
+
+---
+
+## What is an "LLM wiki"?
+
+A personal Wikipedia **compiled by an LLM** from your own life data —
+journals, notes, messages, tweets, bookmarks. Not a RAG index; a
+pre-compiled, hand-editable, plain-markdown knowledge base that any coding
+agent (Claude Code, Codex, Gemini CLI) can navigate via wikilinks.
+
+Three-layer architecture (Karpathy):
+
+1. **Raw** — original sources, untouched (`raw/` or `data/`)
+2. **Wiki** — LLM-compiled markdown articles with `[[wikilinks]]` (`wiki/`)
+3. **Schema** — the `SKILL.md` / `CLAUDE.md` that instructs the LLM how to
+ write and organize
+
+Karpathy's pithy framing: *"Obsidian is the IDE, the LLM is the programmer,
+the wiki is the codebase."*
+
+---
+
+## The tweets that made it viral
+
+- **Andrej Karpathy — original LLM Wiki post (Nov 2025)** — 12M+ views, kicked the whole thing off:
+ https://x.com/karpathy/status/2040572272944324650
+- **Karpathy's gist with the skill/prompt** (2,100+ stars in 12 hours):
+ https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f
+- **Farza Majeed — "Farzapedia"** (the first widely-shared implementation, the
+ reply that made Karpathy's idea concrete for most people):
+ https://x.com/FarzaTV/status/2040563939797504467
+- **Farza's published SKILL.md gist** — the file you drop into
+ `.claude/skills/wiki/SKILL.md` to get started:
+ https://gist.github.com/farzaa/c35ac0cfbeb957788650e36aabea836d
+- **Farza's landing page** (interest gauge): https://farza.com/knowledge
+
+### Write-ups worth reading
+
+- Analytics Vidhya — technical breakdown:
+ https://www.analyticsvidhya.com/blog/2026/04/llm-wiki-by-andrej-karpathy/
+- MindStudio — step-by-step with Claude Code:
+ https://www.mindstudio.ai/blog/andrej-karpathy-llm-wiki-knowledge-base-claude-code
+- Toolmesh deep dive:
+ https://www.toolmesh.ai/news/karpathy-llm-wiki-personal-knowledge-management
+- Medium — what is an LLM Wiki?:
+ https://medium.com/@aristojeff/what-is-an-llm-wiki-and-why-are-people-paying-attention-to-it-b7e10617967d
+
+---
+
+## Ingesting your iMessage history
+
+All your iMessages live in a single SQLite database on every Mac:
+
+**`~/Library/Messages/chat.db`**
+
+Requires **Full Disk Access** for the terminal / app that opens it
+(System Settings → Privacy & Security → Full Disk Access).
+
+Tables worth knowing:
+
+| Table | What's in it |
+|-------|--------------|
+| `message` | Every message — `text`, `attributedBody` (emoji/rich text), `date` (Apple epoch), `is_from_me`, `handle_id` |
+| `handle` | Phone numbers and emails of everyone you've ever texted |
+| `chat` | Conversations (1:1 or group) |
+| `chat_message_join` | Maps messages → chats |
+
+Gotchas:
+
+- `date` is in Apple epoch (nanoseconds since 2001-01-01 UTC). Convert with
+ `datetime(message.date/1e9 + 978307200, 'unixepoch')` in SQL.
+- Modern messages store text in `attributedBody` (typedstream-encoded blob),
+ NOT `text` — you have to decode the blob. The open-source scripts below
+ already handle this.
+- Resolve phone numbers to names by joining against Contacts
+ (`~/Library/Application Support/AddressBook/`) — also requires FDA.
+
+### Skills / scripts that do it for you
+
+The skill file most people use at these events is **Farza's wiki skill**:
+https://gist.github.com/farzaa/c35ac0cfbeb957788650e36aabea836d
+— drop it into `.claude/skills/wiki/SKILL.md` in an empty project, put
+your exports in `data/`, then run `/wiki ingest` and `/wiki absorb all`.
+
+Open-source iMessage ingesters you can crib from:
+
+- **`kothari-nikunj/llm-wiki`** — most complete. Has a dedicated
+ `ingest_imessage.py` that reads `chat.db` directly (both `text` and
+ `attributedBody`), resolves handles to contact names via Address Book,
+ and groups by contact + day. Also supports WhatsApp, Apple Notes,
+ tweets, bookmarks. Ships a Next.js viewer.
+ → https://github.com/kothari-nikunj/llm-wiki
+- **`SamurAIGPT/llm-wiki-agent`** — drop-in agent, works with Claude Code,
+ Codex, or Gemini CLI. Adds NetworkX + Louvain community detection for
+ knowledge-graph visualization.
+ → https://github.com/SamurAIGPT/llm-wiki-agent
+- **`ussumant/llm-wiki-compiler`** — Claude Code plugin focused on the
+ compile step only.
+ → https://github.com/ussumant/llm-wiki-compiler
+- **`nashsu/llm_wiki`** — cross-platform desktop app wrapper.
+ → https://github.com/nashsu/llm_wiki
+
+### Minimal first run
+
+```bash
+mkdir mypedia && cd mypedia
+mkdir -p .claude/skills/wiki
+curl -sL https://gist.githubusercontent.com/farzaa/c35ac0cfbeb957788650e36aabea836d/raw \
+ > .claude/skills/wiki/SKILL.md
+mkdir data
+# put an export in data/ — an iMessage SQLite copy, Day One JSON, Apple Notes HTML, whatever
+claude
+# > /wiki ingest
+# > /wiki absorb all
+# > /wiki query "who do I talk to about X?"
+```
+
+---
+
+## Farzapedia by the numbers (reference)
+
+| | |
+|---|---|
+| Total raw entries ingested | 2,500 |
+| Compiled wiki articles | 400 |
+| Sources | Day One diary, Apple Notes, iMessage |
+| Architecture | Karpathy's 3-layer (raw → wiki → schema) |
+| Agent used to query | Claude Code |
+| Storage | Plain markdown files — no vector DB, no embeddings |
+| Most surprising output | An auto-generated "Relationship History" article (why Farza won't make Farzapedia public) |
+
+---
+
+## Add your own page
+
+This wiki is **public-edit**. Anyone can:
+
+- Edit this page — click *Edit* at the top and save.
+- Create a new page — go to `/@jacobcole/agi-house-llmwiki/new` or `POST`
+ to `/api/v1/wikis/jacobcole/agi-house-llmwiki/pages` (no auth required —
+ anonymous pages are marked claimable).
+- Link to your own LLM wiki — drop it under [Attendee wikis](#attendee-wikis) below.
+
+Pages you might want to add:
+
+- Your attempt, with what worked / what broke
+- A better ingester for a specific source (WhatsApp, Signal, Slack, Email)
+- Tips for the `absorb` / `cleanup` phase
+- Prompts you found useful
+- Trip reports on different models (Opus vs. Sonnet vs. GPT vs. Gemini)
+
+---
+
+## Attendee wikis
+
+*Add yours! `[[name]]` — one-line description — URL*
+
+- *(add your wiki here)*
+
+---
+
+## FAQ
+
+**Do I need to pay for anything?**
+No. `wikihub.md` is free. Claude Code free tier is enough to run a small
+absorb. Farza used Claude.
+
+**Will my private data leak?**
+Depends entirely on where you run the compile. Karpathy's #2 principle is
+**local** — the files stay on your computer. If you point Claude Code at a
+local directory, inference happens via the API but the files never leave
+your machine except in prompts. If you're compiling anything sensitive,
+audit the SKILL.md first and keep the wiki directory out of any cloud
+sync.
+
+**How big does it get?**
+Farza: 2,500 entries → 400 articles, a few MB of markdown. iMessage-only
+ingests can be 10K–100K messages; they compress down to dozens of articles
+because most chats are banal.
+
+**Can I push my wiki back to wikihub.md?**
+Yes — `git clone https://USERNAME:wh_KEY@wikihub.md/@USERNAME/SLUG.git`
+and push. Or use the REST pages API. See `/AGENTS.md` on this host for the
+full agent setup recipe.
+
+---
+
+## Related / background
+
+- **WikiHub home:** https://wikihub.md
+- **WikiHub agent setup:** https://wikihub.md/AGENTS.md
+- **Karpathy's broader take on personal AI:**
+ https://x.com/karpathy/status/2040572272944324650
+- **Ankur Warikoo's "YBAAS" response** (you-are-a-service, one of the
+ better commentaries on why LLM wikis are actually a big deal):
+ https://x.com/warikoo/status/2040595110635901255
+- **IQ Source — "knowledge compounds or rots"** (the case for curation):
+ https://www.iqsource.ai/en/blog/karpathy-llm-wiki-knowledge-compounds-or-rots/
+
+---
+
+*This page is public-editable. Hosted on [WikiHub](https://wikihub.md).
+Created for the AGI House LLM Wiki event, 2026-04-22.*
+
+## Connect your MCP client to WikiHub
+
+Want Claude Desktop, Claude Code, or ChatGPT Deep Research to read and write your WikiHub pages with your own API key? Follow the step-by-step setup guide:
+
+- [[mcp-server-setup|WikiHub MCP Connector — Setup]] — install, configure, deploy
+
+Short version: build `mcp-server/` in the [wikihub repo](https://github.com/tmad4000/wikihub), paste a two-line JSON block into your Claude config, and `wikihub_search` / `wikihub_create_page` / 15 other tools become available inside your agent.
new file mode 100644
index 0000000..c3fe8b6
@@ -0,0 +1,42 @@
+---
+visibility: public-edit
+---
+
+# Links
+
+- [[mcp-server-setup|WikiHub MCP Connector — Setup]] — connect Claude Desktop, Claude Code, or ChatGPT to your WikiHub pages with your own API key. Built as a port of the noos MCP server.
+
+---
+Read my iMessage Chat database and make me a personal wiki in the spirit of the LLM Wiki format. Look that up online ~/Library/Messages/chat.db
+
+Use this skill optionally
+https://gist.github.com/farzaa/c35ac0cfbeb957788650e36aabea836d
+
+OpenMarkdownReader https://github.com/tmad4000/OpenMarkdownReader/releases/tag/v1.0.5
+
+Whatsapp
+- Ask claude: install Matrix Bridge so you can read and export whatsapp data. Also see if you can export whatsapp directly in parallel
+
+
+- Managing your personal network use case. The other use cases we just mentioned were:
+- Managing all the historical Claude sessions
+- managing all the recordings you have personally; - All the granola history that you have
+- Making a wiki out of all the knowledge that is created in AGI House talks
+
+- Managing your health data
+- ingesting existing ChatGPT and Claude.ai normal chats
+In integrating that data
+
+What other sources were we talking about, or what other use cases were we talking about?
+
+
+
+--
+
+## Lists we want to make public
+travel destinations and tips - notion
+restaurants - google maps
+
+list of events - recurring or that I should go to
+great healers / bodyworkers
+list of vcs and angels in my network (share with just friends)
new file mode 100644
index 0000000..361ab57
@@ -0,0 +1,352 @@
+---
+title: WikiHub MCP Connector — Setup
+visibility: public-edit
+tags: [mcp, claude, chatgpt, wikihub, setup]
+---
+
+# WikiHub MCP Connector — Setup
+
+A standalone [Model Context Protocol](https://modelcontextprotocol.io/) server that lets **Claude Desktop**, **Claude Code**, or **ChatGPT** read and write WikiHub pages with your own API key. Ported from the noos MCP server — same structure, same per-request auth isolation, adapted to WikiHub's REST surface.
+
+- **Source:** `github.com/tmad4000/wikihub` → `mcp-server/`
+- **Transports:** stdio (local Claude clients) + Streamable HTTP (ChatGPT / remote)
+- **Auth:** your personal WikiHub API key (`wh_…`), scoped per request
+
+---
+
+## What you get
+
+17 tools + ChatGPT Deep Research aliases:
+
+**Read** (auth optional — public pages work anonymously)
+- `wikihub_search` — fuzzy full-text across pages (scope with `wiki: "owner/slug"`)
+- `wikihub_get_page` — read one page's content + frontmatter
+- `wikihub_list_pages` — list readable pages in a wiki
+- `wikihub_get_wiki` — wiki metadata (title, counts)
+- `wikihub_commit_log` — git history for a wiki
+- `wikihub_shared_with_me` — wikis/pages shared with you
+- `wikihub_whoami` — identity of the current api key
+
+**Write** (auth required, except anonymous posts on `public-edit` wikis)
+- `wikihub_create_wiki`
+- `wikihub_create_page` (pass `anonymous: true` for public-edit anon posts)
+- `wikihub_update_page`
+- `wikihub_append_section` — append under an optional `## heading`, non-destructive
+- `wikihub_delete_page`
+- `wikihub_set_visibility` — `public | public-edit | private | unlisted`
+- `wikihub_share` — grant read/edit to a user or email
+- `wikihub_list_grants`
+- `wikihub_fork_wiki`
+- `wikihub_register_agent` — self-register a new account, returns an `api_key`
+
+**ChatGPT Deep Research**
+- `search`, `fetch` — wrappers over `wikihub_search` / `wikihub_get_page` that use a composite `owner/slug:path` id so `fetch` can resolve the page.
+
+---
+
+## 1. Get an API key
+
+### Option A — curl (one shot, no browser)
+
+```bash
+curl -X POST https://wikihub.md/api/v1/accounts \
+ -H 'Content-Type: application/json' \
+ -d '{"username":"your-agent-name"}'
+```
+
+The response contains `api_key` (starts with `wh_…`). **Save it — it's shown once.**
+
+### Option B — from inside Claude itself
+
+Once the MCP server is installed (even without a key), ask Claude:
+
+> Use `wikihub_register_agent` to sign me up as `my-agent`.
+
+The tool returns an api_key you paste back into your MCP client config.
+
+---
+
+## 2. Build the server
+
+```bash
+git clone https://github.com/tmad4000/wikihub.git
+cd wikihub/mcp-server
+npm install && npm run build
+```
+
+That leaves a runnable `dist/index.js` (stdio) and `dist/http.js` (HTTP).
+
+---
+
+## 3. Connect your client
+
+### Claude Code (stdio — recommended)
+
+```bash
+claude mcp add -s user wikihub -- \
+ env WIKIHUB_API_KEY=wh_yourkey node /absolute/path/to/wikihub/mcp-server/dist/index.js
+```
+
+Restart Claude Code. `wikihub_*` tools show up in the tool picker.
+
+### Claude Desktop
+
+Edit `~/Library/Application Support/Claude/claude_desktop_config.json`:
+
+```json
+{
+ "mcpServers": {
+ "wikihub": {
+ "command": "node",
+ "args": ["/absolute/path/to/wikihub/mcp-server/dist/index.js"],
+ "env": {
+ "WIKIHUB_API_KEY": "wh_yourkey"
+ }
+ }
+ }
+}
+```
+
+Restart the app.
+
+### Claude connector (Claude Desktop / Claude.ai — no local install)
+
+Claude Desktop and Claude.ai both support remote MCP servers via **Settings → Connectors → Add custom connector**. This is the zero-install path — no `npm`, no config file, no node runtime on your machine. A hosted endpoint is live at **`https://mcp.wikihub.md/mcp`** (verify with `curl https://mcp.wikihub.md/healthz` → `{"status":"ok"}`).
+
+1. In Claude (Desktop or claude.ai) open **Settings → Connectors** → **Add custom connector**.
+2. Fill in:
+ - **Name:** `WikiHub`
+ - **Description:** *Read and write your WikiHub pages*
+ - **Remote MCP server URL:** `https://mcp.wikihub.md/mcp`
+3. Advanced settings → **Custom headers**:
+ - Header name: `Authorization`
+ - Header value: `Bearer wh_yourkey`
+
+ If the UI has no custom-header field (some older Claude Desktop builds ship with OAuth-only fields), use the `?key=` fallback instead — change the URL to `https://mcp.wikihub.md/mcp?key=wh_yourkey`. Treat that URL like a password; the key is per-user and revocable.
+4. Save. Start a new chat and the `wikihub_*` tools appear in the tool picker.
+
+What the connector gets you: all 17 `wikihub_*` tools plus the ChatGPT-DR `search`/`fetch` aliases, with your own identity and ACL — so private pages you can read, Claude can read; public-edit wikis you want to post to, Claude can post to.
+
+### ChatGPT (custom connector / Deep Research)
+
+Requires the HTTP transport deployed behind TLS — see **Deploy the HTTP transport** below.
+
+Then in ChatGPT → Settings → Connectors → Custom:
+
+- **URL:** `https://mcp.wikihub.md/mcp`
+- **Auth:** custom header `Authorization: Bearer wh_yourkey`
+
+ChatGPT Deep Research will use the `search` / `fetch` tools automatically.
+
+### Claude Code connected to a hosted server (HTTP)
+
+```bash
+claude mcp add -s user wikihub --transport http \
+ --header "Authorization: Bearer wh_yourkey" \
+ https://mcp.wikihub.md/mcp
+```
+
+---
+
+## 4. Smoke test
+
+```bash
+# List tools (anonymous, no key needed)
+curl -s -X POST http://127.0.0.1:4200/mcp \
+ -H 'Content-Type: application/json' \
+ -H 'Accept: application/json, text/event-stream' \
+ -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'
+
+# Search across public wikis
+curl -s -X POST http://127.0.0.1:4200/mcp \
+ -H 'Content-Type: application/json' \
+ -H 'Accept: application/json, text/event-stream' \
+ -d '{"jsonrpc":"2.0","id":2,"method":"tools/call",
+ "params":{"name":"wikihub_search","arguments":{"query":"hello","limit":3}}}'
+```
+
+For stdio, run `WIKIHUB_API_KEY=wh_… node dist/index.js` and it'll stream MCP protocol on stdin/stdout (logs go to stderr).
+
+---
+
+## Authentication model (explicit)
+
+**The WikiHub MCP server uses plain API-key auth. It does NOT use OAuth or
+client IDs.** This is a deliberate design choice — here's what that means:
+
+### What the server accepts
+
+| Auth path | Where you put the `wh_…` key | Works in |
+| --------------------------------- | ------------------------------------------------ | --------------------- |
+| `Authorization: Bearer <key>` | Client's custom-header field | Claude Code (HTTP), ChatGPT DR, curl |
+| `x-api-key: <key>` | Client's custom-header field | Any client that lets you set arbitrary headers |
+| `?key=<key>` query param in URL | Appended to the connector URL | Claude Desktop custom-connector UI (fallback for builds that only show OAuth fields) |
+| `WIKIHUB_API_KEY` env | Shell / systemd / launch config | stdio transport (local) |
+
+Precedence (first match wins): Bearer → `x-api-key` → `?key=` → env.
+
+### What the server does NOT do
+
+- **No OAuth 2.0 flow.** There is no authorization_endpoint, no token exchange, no refresh token dance.
+- **No client ID / client secret.** You don't register your app ahead of time.
+- **No OpenID Connect / OIDC.** The user is identified solely by which `wh_…` key they present.
+- **No OAuth discovery endpoints.** `/.well-known/oauth-authorization-server`, `/.well-known/oauth-protected-resource`, and `/.well-known/openid-configuration` all return `404`. If a client tries OAuth-first discovery, it will fail over to the custom-header / query-param path.
+
+### Why no OAuth
+
+- **One-shot agent onboarding is a core WikiHub principle** (see `AGENTS.md §1`). `POST /api/v1/accounts` returns an immediately-usable `api_key` with no email step, no browser, no consent screen. OAuth would reintroduce all three.
+- **Keys are per-user, labeled, and revocable.** You can mint one key per agent (`POST /api/v1/keys {"label":"claude-connector"}`) and revoke in isolation — fine-grained enough for the current trust model without OAuth scoping.
+- **Every major MCP client already speaks Bearer.** Claude Code's `--header`, ChatGPT Deep Research's custom-header field, and stdio env vars all carry a raw key just fine.
+- **The one client UI that can't paste headers — older Claude Desktop builds — is handled by the `?key=` query-param fallback.** That's a pragmatic tradeoff (URL-visibility risk in exchange for zero-friction install) and keys are revocable if leaked.
+
+### When OAuth would be worth adding
+
+OAuth becomes the right tool when:
+
+1. You want **third-party apps** (not just the user's own agent) to request scoped access without the user pasting a key.
+2. You want to publish the connector in Anthropic's or OpenAI's curated connector directory — some directories require OAuth for discoverability.
+3. You want **per-session scopes** (e.g. "read-only", "one specific wiki", short-lived tokens).
+4. You want a **browser consent screen** for UX reasons — users who don't know what a Bearer token is.
+
+None of that is required for the current "bring your own agent" flow. If we hit one of these use cases later, the upgrade path is additive: keep the `wh_…` Bearer path for backward compat, bolt on an OAuth server that issues short-lived access tokens.
+
+---
+
+## Env vars
+
+| Var | Default | Notes |
+| ----------------- | ---------------------- | ----------------------------------------------------- |
+| `WIKIHUB_API_URL` | `https://wikihub.md` | Override for local dev (e.g. `http://localhost:5100`) |
+| `WIKIHUB_API_KEY` | _unset_ | Required for writes and private reads |
+| `PORT` | `4200` | HTTP transport listen port |
+| `HOST` | `0.0.0.0` | HTTP transport listen interface |
+
+### HTTP auth header precedence
+
+1. `Authorization: Bearer <key>` — preferred
+2. `x-api-key: <key>` — convenience
+3. `?key=<key>` query param — Claude Desktop workaround (its custom-connector UI has no custom-header field as of early 2026)
+4. `WIKIHUB_API_KEY` env — local / single-tenant fallback
+
+Each HTTP request builds a **fresh `McpServer` closed over its own api client**, so two concurrent callers can't leak keys into each other's sessions.
+
+---
+
+## Deploy the HTTP transport
+
+Sibling to the main app on Lightsail `wikihub-dev`:
+
+```bash
+ssh -i ~/.ssh/wikihub-dev-key ubuntu@54.145.123.7 \
+ 'cd /opt/wikihub-app/mcp-server && git pull && npm install && npm run build && sudo systemctl restart wikihub-mcp'
+```
+
+Suggested systemd unit (`/etc/systemd/system/wikihub-mcp.service`):
+
+```ini
+[Unit]
+Description=WikiHub MCP server (Streamable HTTP)
+After=network.target
+
+[Service]
+Type=simple
+User=ubuntu
+WorkingDirectory=/opt/wikihub-app/mcp-server
+Environment=PORT=4200
+Environment=HOST=127.0.0.1
+Environment=WIKIHUB_API_URL=https://wikihub.md
+ExecStart=/usr/bin/node dist/http.js
+Restart=on-failure
+
+[Install]
+WantedBy=multi-user.target
+```
+
+Nginx in front on `mcp.wikihub.md`:
+
+```nginx
+server {
+ listen 443 ssl http2;
+ server_name mcp.wikihub.md;
+ location / {
+ proxy_pass http://127.0.0.1:4200;
+ proxy_http_version 1.1;
+ proxy_set_header Host $host;
+ proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+ proxy_read_timeout 300;
+ }
+}
+```
+
+Cloudflare DNS: A record `mcp.wikihub.md → 54.145.123.7` (proxied).
+
+---
+
+## Design notes (ported from noos)
+
+- **Same module layout** as `noos/mcp-server/src/`: `api.ts`, `server.ts`, `instructions.ts`, `index.ts` (stdio), `http.ts` (HTTP).
+- **`User-Agent: curl/8.0`** is the default on all outbound requests — Cloudflare blocks some non-curl UAs in front of `wikihub.md`.
+- **Provenance headers** `X-Agent-Name: wikihub-mcp` and `X-Agent-Version: …` land in the server's `ApiKey.agent_name` / `agent_version` columns at key lookup, so every write is traceable to the connector.
+- **Personalized instructions** — on connect, the server calls `whoami` and seeds the MCP `instructions` string with the caller's identity so the model knows who it is from turn one.
+
+---
+
+## Troubleshooting
+
+- **"WIKIHUB_API_KEY is not set"** — writes need a key even though reads don't. Pass it in the client config, via `Authorization: Bearer` on HTTP, or as the env var.
+- **Cloudflare 403** — the default `curl/8.0` UA should get through; if you override `userAgent`, pick something Cloudflare likes.
+- **`tools/list` returns empty** — check stderr of the stdio process; the server logs `[wikihub-mcp] v0.1.0 → <url> (authenticated|anonymous)` on startup.
+- **ChatGPT DR can't find the connector** — make sure the URL ends in `/mcp` and your header is `Authorization: Bearer wh_…` (not `x-api-key`; DR only speaks Bearer).
+
+---
+
+## Skill — `/wikihub-build` (Farza's `/wiki` adapted to WikiHub)
+
+Once the connector is loaded in your client, install the companion **skill** — a WikiHub-native port of [Farza Majeed's canonical `/wiki` skill](https://gist.github.com/farzaa/c35ac0cfbeb957788650e36aabea836d) that the AGI House LLM Wiki event is centered on. Same three-layer Karpathy pattern, same writing standards, but writes to two hosted WikiHub wikis (`personal-wiki-raw` + `personal-wiki`) via MCP instead of local folders — so your wiki lives at a URL, survives machine hops, and you can share selected articles per-page.
+
+One-line install:
+
+```bash
+mkdir -p ~/.claude/skills/wikihub-build
+curl -fsSL https://raw.githubusercontent.com/tmad4000/wikihub/main/skills/wikihub-build/SKILL.md \
+ > ~/.claude/skills/wikihub-build/SKILL.md
+```
+
+Invoke inside Claude Code:
+
+```
+/wikihub-build ingest # Parse your source data into @you/personal-wiki-raw
+/wikihub-build absorb all # Compile raw entries into articles in @you/personal-wiki
+/wikihub-build query <q> # Ask questions
+/wikihub-build cleanup # Audit and enrich existing articles (parallel subagents)
+/wikihub-build breakdown # Find and create missing articles
+/wikihub-build status
+```
+
+Claude Desktop / Claude.ai discover it when you mention the skill name in chat. The MCP connector above must be loaded in the same client — the skill is the orchestration, the connector provides the tools.
+
+Canonical source: [`skills/wikihub-build/SKILL.md`](https://github.com/tmad4000/wikihub/blob/main/skills/wikihub-build/SKILL.md) in the WikiHub repo.
+
+### vs. Farza's original
+
+| | Farza `/wiki` (local files) | WikiHub `/wikihub-build` (hosted) |
+|-----------------|------------------------------------|---------------------------------------------|
+| Storage | `raw/` + `wiki/` on disk | Two private WikiHub wikis, URL-addressable |
+| Backlinks | `_backlinks.json` rebuilt manually | Native WikiHub `Wikilink` model — automatic |
+| Sharing | Whole-repo (git push) | Per-page `wikihub_set_visibility` + ACL |
+| Multi-machine | Diverges | Single source of truth |
+| Portability | Copy the folder | Every page has a stable HTTPS URL |
+| Publish a page | git publish | Flip visibility, no git needed |
+
+Both are valid. Pick Farza's if you want Obsidian-as-IDE locally, WikiHub-build if you want the wiki to be an internet-citizen your other agents can reach.
+
+---
+
+## Related on this wiki
+
+- [[index|AGI House — LLM Wiki Event]] — the event this wiki is for
+- [[links|Links]] — tweets, write-ups, and other resources people have dropped here
+
+---
+
+*Last updated 2026-04-22 — matches `mcp-server/` at commit `4defc58` on the `cli-installer` branch.*