You are helping me build the following AI agent workflow.

## Goal
Nutrition And Fitness Tracker: Users log food or activity via webhook; the system estimates calories, updates daily totals in Google Sheets.

## Specification
- What it does: You automatically log your meals and workouts via a webhook to instantly get calorie and protein estimates while tracking your daily totals in Google Sheets. You receive immediate Slack alerts whenever you exceed your limits or face potential health risks, ensuring you stay informed about your nutrition and fitness progress without manual entry.
- Trigger: Triggered by an event (Webhook input and Scheduled daily reset)
- Autonomy: Fully hands-off
- Expected setup effort: about an afternoon
- Tools/services involved:
  - OpenAI Chat Model: Classify food or activity and estimate nutrition data
  - Google Sheets: Store user profiles and daily metrics
  - Slack: Send alerts for calorie limits or health risks

## Known pitfalls, handle each one explicitly in your implementation
1. Ambiguous user inputs like 'lunch' can confuse the classifier, so add a system prompt that requests specific details or defaults to a generic entry.
2. Google Sheets API quotas may be exceeded during high-volume testing, so implement batch updates instead of row-by-row writes.
3. Slack alerts might spam users if thresholds are too sensitive, so add a cooldown period or daily summary aggregation for non-critical warnings.
4. Calorie estimates from LLMs can vary significantly between runs, so standardize the output format and consider adding a confidence score to each prediction.

## Reference implementation
https://n8n.io/workflows/14858 (workflow template)
Fetch and inspect this before building. If it matches my stack, adapt it;
if not, rebuild the pattern with my tools.

## Process requirements
1. Before building: ask me which of the listed tools I actually use and
   what my platform is (n8n / Make / code / other). Do not assume.
2. Adapt the pattern to my answers; do not force the reference stack.
3. Address every pitfall above; tell me how you handled each.
4. Provide a test plan I can run before letting this touch real data.
5. Ask before any step that sends messages, modifies data, or spends money.

Source: https://usecasesforagents.com/use-case/automated-nutrition-and-fitness-tracker/ via usecasesforagents.com