First Agent Template — Local Run Guide

This file explains how to run the prototype locally (macOS, zsh).

Assumptions
- Project root: /Users/gautam.kumar/Desktop/agents/First_agent_template
- You will use the included virtualenv `.venv` or create one.

1) Create & activate a venv (if needed)

```bash
cd /Users/gautam.kumar/Desktop/agents/First_agent_template
python3 -m venv .venv
source .venv/bin/activate
```

2) Install dependencies

```bash
pip install --upgrade pip
pip install -r requirements.txt
# optional: install gradio UI extras
pip install "smolagents[gradio]"
```

3) Smoke tests

```bash
PYTHONPATH=$(pwd) python tests/run_nlu_test.py
```

4) Demo run

```bash
PYTHONPATH=$(pwd) python tests/run_dialog_demo.py
```

5) Interactive CLI

```bash
PYTHONPATH=$(pwd) python tools/interactive_cli.py
```

Interactive CLI tips
- Type natural sentences like: "I can pay ¥30000 by next Friday"
- If assistant asks for missing info ("Can I confirm your full name?"), reply with the requested info.
- To request a human operator: "I want to talk to an operator tomorrow morning" — the assistant will propose JST slots.
- Type `exit` or `quit` to leave.

Data & files
- `data/requests.json` — saved handoff requests (JSON).
- `data/operator_notifications.log` — operator notification stub logs.

If you want, I can:
- Add this file into the main `README.md` (merge), or keep it separate.
- Add unit tests and CI configs.
- Improve the CLI to track dialogue state across multiple turns more robustly.