Skip to content

README.md

Latest commit

 

History

History
94 lines (66 loc) · 2 KB

README.md

File metadata and controls

94 lines (66 loc) · 2 KB

Examples

Integration examples for the Prompt Injection Firewall (PIF). Each example demonstrates how to route LLM API requests through PIF for real-time prompt injection detection.

Prerequisites

  1. PIF proxy running:

    # For OpenAI
pif proxy --target https://api.openai.com --listen :8080

# For Anthropic
pif proxy --target https://api.anthropic.com --listen :8080

  2. API key for your LLM provider (OpenAI or Anthropic).

Examples

Directory Language Description
python/ Python OpenAI and Anthropic SDK integration
nodejs/ Node.js OpenAI SDK integration with async/await
curl/ Shell Raw HTTP requests for testing
docker/ Docker Production-ready Docker Compose setup

Python

cd python
pip install -r requirements.txt

# OpenAI example
OPENAI_API_KEY=sk-... python openai_example.py

# Anthropic example
ANTHROPIC_API_KEY=sk-ant-... python anthropic_example.py

Node.js

cd nodejs
npm install

OPENAI_API_KEY=sk-... node openai_example.js

cURL

cd curl

# OpenAI
OPENAI_API_KEY=sk-... bash openai.sh

# Anthropic
ANTHROPIC_API_KEY=sk-ant-... bash anthropic.sh

Docker

cd docker
docker compose up -d

# Verify PIF is running
curl http://localhost:8080/healthz

# Then point your SDK at http://localhost:8080/v1

What Each Example Demonstrates

Every example shows three scenarios:

  1. Clean prompt -- A benign request that passes through PIF to the LLM API
  2. Prompt injection -- An attempt to override system instructions (blocked with HTTP 403)
  3. Data exfiltration / jailbreak -- An attempt to extract data or bypass safety (blocked with HTTP 403)

Expected Output

When PIF blocks a request, you will receive:

{
  "error": {
    "message": "Request blocked by Prompt Injection Firewall",
    "type": "prompt_injection_detected",
    "score": 0.85,
    "findings": 2
  }
}