superagent

Overview of superagent

What it is

Superagent is an open-source SDK for AI agent safety. It protects AI applications against prompt injections, data leaks, and harmful outputs. The SDK allows developers to embed safety directly into their applications and prove compliance to customers. It is licensed under MIT and available as open source.

Key features

• Guard: Detect and block prompt injections, malicious instructions, and unsafe tool calls at runtime.
• Redact: Automatically remove PII, PHI, and secrets from text.
• Scan: Analyze repositories for AI agent-targeted attacks such as repo poisoning and malicious instructions.
• Test: Run red team scenarios against a production agent (coming soon).
• Works with any model (OpenAI, Anthropic, Google, Groq, Bedrock, and more).
• Open-weight models allow running Guard on your infrastructure with 50-100ms latency.
• Low latency optimized for runtime use.

Use cases

Superagent can be used to secure AI applications by blocking prompt injections, redacting sensitive data from user inputs or outputs, scanning code repositories for threats, and testing agents against red team scenarios. It is suitable for any application that uses AI models and needs to enforce safety policies.

Getting started

Sign up at superagent.sh to get an API key. Install the SDK: bash npm install safety-agent bash uv add safety-agent Set your API key: bash export SUPERAGENT_API_KEY=your-key

License and pricing

The license is MIT. The project is open source with no stated pricing.

Based on the project's public README and repository metadata.