Superagent

Stale
GitHub TypeScript MIT

Description

Superagent protects AI applications against prompt injections, data leaks, and harmful outputs, embedding safety directly into your app.

Key Features

  • Guard: detect and block prompt injection attacks at runtime
  • Redact: automatically remove PII, PHI, and secrets from text
  • Scan: analyze repositories for AI agent-targeted security threats
  • TypeScript and Python SDKs plus CLI tools available
  • Open-weight models run locally with 50-100ms latency
  • Works with OpenAI, Anthropic, Google, and all major model providers

Use Cases

💡 Deploy prompt injection protection at the AI application entry point
💡 Automatically redact sensitive personal information from user inputs
💡 Scan code repositories to detect supply chain attacks and malicious instructions
💡 Run red team scenarios against production AI agents

Quick Start

1. Sign up at superagent.sh to get your API key
2. TypeScript: `npm install safety-agent` or Python: `uv add safety-agent`
3. Set environment variable `SUPERAGENT_API_KEY`
4. Call `client.guard()` for security detection
5. Use `client.redact()` to redact sensitive data

Related Projects