gpt_oss_safety_agent

GPT-OSS Safety Agent Example with OpenRouter Provider

A simple example demonstrating OpenAI's gpt-oss-safeguard-20b model via OpenRouter provider with PIIBlockPolicy for detecting and blocking personally identifiable information.

Prerequisites

Set your OpenRouter API key:

export OPENROUTER_API_KEY="your-api-key"
export OPENAI_API_KEY="your-api-key"

Installation

Install dependencies from upsonic_configs.json:

upsonic install

Managing Dependencies

Add a new package:

upsonic add <package> <section>
# Examples:
upsonic add requests api
upsonic add pandas==2.0.0 api

Remove a package:

upsonic remove <package> <section>
# Examples:
upsonic remove requests api

Sections: api, streamlit, development

Run Options

Option 1: Direct Script

uv run main.py

Runs built-in test cases (safe query + PII query).

Option 2: FastAPI Server

upsonic run

Starts server at http://localhost:8000 with Swagger docs at /docs.

Example API call:

curl -X POST http://localhost:8000/call \
  -H "Content-Type: application/json" \
  -d '{"user_query": "My email is john@example.com, can you help me with my account?"}'

How It Works

• Safe queries → Normal AI response
• Queries with PII (emails, phone numbers, etc.) → Blocked with helpful feedback suggesting how to rephrase

Key Files

File	Purpose
main.py	Agent with safety policy
upsonic_configs.json	Dependencies & API schema