nateherkai

Advisor Tool Demo — Development skill for Claude Code

Development community

Demo app for Anthropic's Advisor Tool (beta) — customer support agent comparing cost/quality tradeoffs across Haiku, Sonnet, and Opus configurations.

How to install Advisor Tool Demo

This entry records only its repository, not the path inside it, so there is no exact command to give. Open nateherkai/advisor-tool-demo and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Advisor Tool Demo does

Demo app for Anthropic's Advisor Tool (beta) — customer support agent comparing cost/quality tradeoffs across Haiku, Sonnet, and Opus configurations.

Alternatives in Development

Ideal Customer Profile — Identify ICP with demographics, behaviors, and JTBD 7.8k ★
Crypto Research Haiku Command — Execute lightweight cryptocurrency research by calling only haiku versions of crypto analysis agents in parall 3.4k ★
Claude Keysmith — Managed Claude Code instruction deployment with safe CLI recovery and an unsigned macOS/Windows desktop beta 696 ★

README

Advisor Tool Demo

A demo app showcasing Anthropic's [Advisor Tool](https://docs.anthropic.com/en/docs/build-with-claude/advisor) (beta). A customer support agent for a fictional company "TechFlow" that compares cost and quality tradeoffs across different model configurations.

What It Does

The app runs a customer support chatbot with an agentic tool loop. It has four modes you can switch between:

Mode	Executor	Advisor	Use Case
Haiku + Opus Advisor	Haiku 4.5	Opus 4.6	Cheap executor, smart advisor on complex queries
Sonnet Solo	Sonnet 4.6	None	Baseline — one model does everything
Sonnet + Opus Advisor	Sonnet 4.6	Opus 4.6	Mid-tier executor with advisor backup
Opus Solo	Opus 4.6	None	Maximum quality, highest cost

The real-time dashboard on the right side of the screen shows every model call, token counts, and cost breakdowns — color-coded by model.

Setup

**Prerequisites:** Python 3.10+ and an [Anthropic API key](https://console.anthropic.com/).

# Clone the repo
git clone https://github.com/nateh/advisor-tool-demo.git
cd advisor-tool-demo

# Create a virtual environment
python -m venv venv
source venv/bin/activate        # macOS/Linux
# venv\Scripts\activate          # Windows

# Install dependencies
pip install -r requirements.txt

# Add your API key
cp .env.example .env
# Edit .env and paste your Anthropic API key

# Run
python app.py

Open [http://localhost:8000](http://localhost:8000) in your browser.

How It Works

The backend (`app.py`) runs an agentic loop:

Sends the user message to Claude with tools (knowledge base search, ticket creation, escalation)
In advisor modes, also includes the advisor tool — backed by Opus 4.6
Executes any tool calls locally, sends results back
Advisor calls are handled server-side by the API (no local execution needed)
Loops until the model is done responding
Returns the response pl