Phase 3b: FastMCP "ai service" for claude.ai cooking mode

A standalone MCP server (no Django import) exposing 7 tools over
Streamable HTTP, backed by the Django REST API via the caine token.

- mcp_server/: client.py (httpx wrapper over /api/), server.py (the
  tools + FastMCP app + main), __main__.py, tests.py.
- Tools: get_pantry, set_item_state, add_to_pantry, what_can_i_cook,
  get_recipes, log_cook (suggests, never mutates), create_meta_recipe
  (brainstorm -> commit). Each description says when to call it.
- deploy/food-mcp.service: systemd unit (own process, runs .venv python
  -m mcp_server, loads /var/lib/food/.env).
- deploy/food.tomflux.xyz.nginx: current config + an authless
  /mcp/<secret>/ location proxying to 127.0.0.1:8765 (the URL secret is
  the credential; SSE-friendly buffering/timeout).
- pyproject: [dependency-groups] mcp = [fastmcp, httpx]; deploy with
  `uv sync --group mcp`.

Verified: 7 tools register on fastmcp 3.x, run() accepts transport/
host/port/path, 6 FoodClient unit tests pass (httpx MockTransport).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

1 files changed, 14 insertions, 8 deletions

diff --git a/mcp.md b/mcp.md
index 06449fb..42a491d 100644
--- a/mcp.md
+++ b/mcp.md
@@ -8,7 +8,7 @@
 > Decisions locked in `plan.md §3` and `requirements.md §7`: a **separate
 > FastMCP "ai service"** (a pattern Tom already runs elsewhere) that talks to
 > the existing Django REST API, hosted on the same box at
-> **`food.jihakuz.xyz/mcp`**, behind nginx, **bearer-token** auth.
+> **`food.tomflux.xyz/mcp`**, behind nginx, **bearer-token** auth.
 
 ---
 
@@ -43,7 +43,7 @@ blocks — now grounded instead of guessed.
   claude.ai (cooking mode)
         │  Streamable HTTP, secret in the URL path
         ▼
-  nginx  food.jihakuz.xyz/mcp   (TLS, public)
+  nginx  food.tomflux.xyz/mcp   (TLS, public)
         │  proxy_pass 127.0.0.1:8765
         ▼
   FastMCP "ai service"  (food-mcp.service, localhost only)
@@ -71,7 +71,7 @@ localhost Django**. It holds no database connection of its own.
 
 - **Transport:** Streamable HTTP (the remote-MCP transport claude.ai connectors
   use). FastMCP served over HTTP, bound to `127.0.0.1:8765`.
-- **Public route:** nginx on `food.jihakuz.xyz`, `location /mcp` →
+- **Public route:** nginx on `food.tomflux.xyz`, `location /mcp` →
   `proxy_pass http://127.0.0.1:8765`. TLS via the existing Let's Encrypt setup.
   Note SSE/streaming needs `proxy_buffering off;` and a long
   `proxy_read_timeout` on that location.
@@ -90,7 +90,7 @@ header**, so the original bearer-header plan isn't directly supported. Two ways
 to honour the "tight secret, single user" intent:
 
 - **Recommended — authless connector + secret in the URL path.** Register the
-  connector URL as `https://food.jihakuz.xyz/mcp/<long-random-secret>/`. To
+  connector URL as `https://food.tomflux.xyz/mcp/<long-random-secret>/`. To
   claude.ai it's an authless server; in reality nginx proxies *only* that exact
   secret prefix to the FastMCP service and 404s everything else. The URL **is**
   the bearer-equivalent — it matches Tom's "a tight enough secret, I'm the only
@@ -178,11 +178,17 @@ A small package in this repo, its own process — not bolted into Django:
 ```
 mcp_server/
   __init__.py
-  __main__.py        # FastMCP app; runs streamable-HTTP on 127.0.0.1:8765
+  __main__.py        # entrypoint: from .server import main; main()
+  server.py          # FastMCP app + the @mcp.tool definitions (§5) + main()
   client.py          # thin httpx client around the Django API (caine token)
-  tools.py           # the @mcp.tool definitions from §5
+  tests.py           # FoodClient tests via httpx.MockTransport (no network)
 ```
 
+`main()` runs streamable-HTTP at path `/` on `127.0.0.1:8765`
+(`mcp.run(transport="http", host, port, path="/")`); nginx maps
+`/mcp/<secret>/` → that root. Client tests run with
+`python -m unittest mcp_server.tests` (needs the `mcp` dep group).
+
 - Deps via a uv group so they only install where needed:
   `[dependency-groups] mcp = ["fastmcp", "httpx"]`. Deploy with
   `uv sync --group mcp` on the box that runs the service.
@@ -204,10 +210,10 @@ All out of git; the systemd unit loads the same `.env` the Django service uses.
 1. `uv sync --group mcp` on the box.
 2. Add the three env vars to `/var/lib/food/.env`; generate the bearer token.
 3. Install `deploy/food-mcp.service`, `daemon-reload`, start, enable.
-4. Add the `location /mcp` block to the `food.jihakuz.xyz` nginx config
+4. Add the `location /mcp` block to the `food.tomflux.xyz` nginx config
    (`proxy_buffering off`, long read timeout), `nginx -t`, reload.
 5. In claude.ai, add a custom connector pointing at the **secret URL**
-   `https://food.jihakuz.xyz/mcp/<secret>/` (authless connector — see §4).
+   `https://food.tomflux.xyz/mcp/<secret>/` (authless connector — see §4).
 6. Smoke test: in cooking mode, "what's in my pantry?" → Claude calls
    `get_pantry` and lists real items.