[paultristanwagner]

All code in this PR was written by an AI agent (Claude Opus 5.5) working with me; I directed it and tested the results on my own account. Feel free to take, change or drop any part of it.

Part of #73.

Adds a set_macro_split tool, and get_macro_targets now reports the
current split. On accounts without macro schedules or templates (the
free tier), get_macro_targets returned two empty lists.

The app stores the split as three profile prefs (percent of energy). It
changes them with POST /api/v2/set_pref, one call per changed macro:

{"key": "macroProtein", "val": "30.0", "auth": {…}, "config": {"call_version": 1}}

The keys are macroProtein, macroCarbs and macroLipids. Current values
are in get_profile → prefs, a list of one-key objects. Verified against
the live API: a split written this way reads back unchanged, and the
per-nutrient targets (GET /api/v3/user/{id}/targets) were unchanged
afterwards.

Changes:

• client: get_macro_split() reads the prefs.
  set_macro_split(protein, carbs, fat) checks that each share is 0-100
  and that they add up to 100, then writes all three.
• server: set_macro_split tool. Its description tells the model to
  confirm with the user and to read the old split first. get_macro_targets
  gains a macro_split field.
• README: tool table rows.

https://claude.ai/code/session_01MCJFDo71qoSqTRtcGfkpXJ