IRLFirst physical meetup — Bengaluru, Sat May 23, 4PM · RSVP on Luma
HomeToolsMCPHow It WorksStoriesPhilosophyCommunityArchitectureStar on GitHub
All Tools
S
OtherFreeOpen Source

SUPERAGENT

Make your AI apps safe with open-source guardrails and red teaming

MIT

ABOUT

Modern AI apps rely on agents that read data, follow instructions, call tools, and produce outputs users trust. These systems fail in ways traditional software does not—prompt injections, data leakage, unsafe tool actions, workflow failures, and compliance violations. Superagent provides open-source guardrails that detect and block prompt injections, redact sensitive information, scan codebases for agent-targeted attacks, and enable red-team testing so developers can embed safety directly into their apps and prove compliance to customers.

INSTALL
pip install safety-agent npm install safety-agent

INTEGRATION GUIDE

1. Detect and block prompt injections, malicious instructions, and unsafe tool calls at runtime 2. Automatically redact PII, PHI, and secrets from agent inputs and outputs 3. Scan repositories for AI agent-targeted attacks such as repo poisoning and malicious instructions 4. Run automated red-team scenarios against production agents to find vulnerabilities before attackers do 5. Secure RAG pipelines and image uploads against adversarial inputs

TAGS

securityguardrailsprompt-injectionai-safetyred-teampii-redaction
Superagent — AI Tool | Agentic AI For Good