#1313th of 14 in Security
superagent
SDK that blocks prompt injections and redacts PII in AI agent apps
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 6.8k
- License
- MIT
- Last commit
- Aug 2026
- Last release
- Sep 2025
- Language
- TypeScript
6,464 on 4 Apr 20266,787 on 31 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Superagent is a TypeScript and Python SDK (package `safety-agent`) with a guard call for detecting prompt injections and unsafe tool calls, a redact call for removing PII, PHI and secrets, and a scan call for finding repo poisoning aimed at agents. A red-team `test` call is listed as coming soon. It needs an API key from superagent.sh, and also ships a CLI and an MCP server.
Who it is for: Developers adding prompt-injection and PII protection to LLM apps
Strengths
- Guard, redact and scan exposed as simple calls in both TypeScript and Python
- Works with several providers: OpenAI, Anthropic, Google, Groq, Bedrock
- Open-weight Guard models on HuggingFace, quoted at 50-100ms latency when self-hosted
- Also ships a CLI and an MCP server for Claude Code and Claude Desktop
Weaknesses
- Setup requires an API key from the hosted superagent.sh service
- Red-team `test` feature is marked coming soon, not available yet
- No self-hosting instructions or Docker setup in the README
- Latest release is 2025-09-14, older than recent commits
What it needs
- Needs Superagent API key (superagent.sh)
- Models: OpenAI, Anthropic, Google, Groq, Bedrock
Quick answers
Is superagent open source?
Yes. Its code is on GitHub under the MIT license.
Also in Security
See all 14| Rank | Project | Score |
|---|---|---|
| 1 | shannonAutonomous pentester that exploits web apps and APIs using source code analysis | 91 out of 100 |
| 2 | ifixaiPython CLI that audits AI agents for policy, access and reliability failures | 86 out of 100 |
| 3 | StrixAutonomous AI pentesting agents that validate vulnerabilities with working exploits | 84 out of 100 |
| #4 | pentagiAutonomous multi-agent penetration testing system with sandboxed security tools | 78 out of 100 |
| #5 | ai-infra-guardRed-teaming platform for AI infra CVEs, MCP servers, agent skills and jailbreaks | 71 out of 100 |