Acacian/aegis

Overview

Acacian/aegis is an LLM guardrails and prompt injection detection library for Python. It auto-instruments LangChain, CrewAI, OpenAI, LiteLLM, and eight other frameworks with a single line of code, providing PII masking, toxicity detection, and policy CI/CD without code changes.

Best for

Best for
Python developers who need to add basic safety controls to existing LLM applications quickly

Use cases

• Integrate prompt injection detection into LangChain agents
• Add PII masking to OpenAI API calls
• Automate policy enforcement for LLM outputs in CI/CD

How to use

Install

pip install agent-aegis

Tested with

Claude Desktop, Cursor, Windsurf, VS Code, ChatGPT

Notes

Acacian/aegis is an LLM guardrails and prompt injection detection library for Python. It auto-instruments LangChain, CrewAI, OpenAI, LiteLLM, and eight other frameworks with a single line of code, providing PII masking, toxicity detection, and policy CI/CD without code changes.

9 stars on GitHub. Last updated 2026-05-05. Licensed MIT.

Use cases

• Integrate prompt injection detection into LangChain agents
• Add PII masking to OpenAI API calls
• Automate policy enforcement for LLM outputs in CI/CD

Pros

• Minimal integration effort with one-line setup
• Supports a wide range of popular LLM frameworks
• Combines multiple guardrails (injection, PII, toxicity) out of the box

Cons

• Python-only, limiting use in polyglot stacks
• May introduce latency overhead during inference
• Requires familiarity with guardrail configuration for custom policies

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.