# Connect your agent to Jenks

> A five-minute tutorial that takes you from nothing to an AI agent answering a question about Jenks Guo, by web, by MCP or by CLI.

## What you will build

By the end of this tutorial your AI agent will have read Jenks Guo's brief and answered a real
question about his experience. Pick one of three paths. Each takes about five minutes.

You need either a web chat assistant (ChatGPT, Claude or Perplexity), Claude Code, or a terminal
on macOS or Linux.

## Path A — a web assistant, no setup

1. Open `https://jenksguo.xyz/agents`.
2. Click **Copy for your agent**. This copies a short prompt that tells an assistant where Jenks's
   brief, `llms.txt` and corpus live, and how to use them.
3. Paste it into ChatGPT, Claude or another assistant that can browse the web, and send it.
4. Or click **Open in ChatGPT** (or Claude, or Perplexity) to start a chat with the prompt already
   filled in.
5. Ask a follow-up, for example: *"Which of Jenks's roles best show solution architecture
   experience?"*

You will see the assistant read `https://jenksguo.xyz/llms.txt` and `https://jenksguo.xyz/brief.md`
and answer with links to pages such as `/career/ericsson-integration-engineer`.

## Path B — Claude Code with the remote MCP server

1. Add the server:

   ```bash
   claude mcp add --transport http jenks https://jenksguo.xyz/mcp
   ```

2. Start Claude Code and check the server is connected:

   ```bash
   claude
   /mcp
   ```

   You should see `jenks` with tools such as `get_brief`, `list_experience`, `get_entry`,
   `list_lenses`, `search_jenks`, `get_document` and `ask_jenks`.

3. Try three prompts, one per tool:

   - *"Use get_brief and summarise Jenks in three bullets."*
   - *"Use list_experience with the lens solution-architect and list the roles."*
   - *"Use ask_jenks: is Jenks a fit for a Head of AI role? Show the evidence and the gaps."*

Notice that `ask_jenks` answers in the third person, separates what Jenks **has done** from what he
**could do**, and states gaps plainly. That is deliberate — see
[Grounding and honesty](/agents/docs/explanation/grounding-and-honesty).

## Path C — the `jenks` CLI

1. Install it:

   ```bash
   curl -fsSL https://jenksguo.xyz/downloads/install.sh | sh
   ```

   The installer downloads the right binary for your platform. Windows users can download
   `jenks-windows-x64.exe` from `https://jenksguo.xyz/agents`, and anyone with Node 18+ can run
   `node jenks.mjs` instead.

2. Read the brief:

   ```bash
   jenks brief
   ```

3. Look at Jenks through a role lens:

   ```bash
   jenks list --lens solution-architect
   ```

4. Ask the assistant:

   ```bash
   jenks ask "Is Jenks a fit for a Head of AI role?"
   ```

5. Open one entry in full:

   ```bash
   jenks get xero-developer-evangelist
   ```

## What you learned

- The corpus is the same everywhere: the website, the files, the API, MCP and the CLI.
- `get_brief` is the best first call; `list_experience` and `get_entry` give the detail;
  `ask_jenks` reasons over everything.
- The same tools work in any language the site supports: add `--locale zh` (CLI) or
  `"locale": "zh"` (MCP) for Simplified Chinese, or `zh-hant`, `ja`, `fr`, `es`, `eo`.

## Next steps

- Add the server to other clients: [Add MCP to your client](/agents/docs/how-to/add-mcp-to-your-client).
- Run a structured role-fit check: [Evaluate role fit](/agents/docs/how-to/evaluate-role-fit).
- See every tool and parameter: [MCP tools reference](/agents/docs/reference/mcp-tools).