AI guardrails & agent security
3,625 agents · most starred first
AI guardrails & agent security, page 24 of 37
A package for LLM security features
Cognitive Security Middleware - The 'Electronic Stability Program' (ESP) for Large Language Models. Bidirectional containment system with defense-in-depth architecture (6 validation layers), stateful tracking, and mathematical safety constraints. Validated against Unicode/encoding attacks, pattern evasion, multilingual/polyglot attacks (12+ languages including Basque, Maltese), and memory/session attacks. Protocol-based hexagonal architecture with LangChain integration.
Runtime monitoring SDK for AI applications — detect prompt injections and adversarial attacks in production.
Safety and guardrails toolkit for LLM applications
Provider and client helpers for signed LLM transcript verification
SIREN: lightweight, plug-and-play guard models reading LLM internal representations, for harmful content and for agent trajectories.
Streaming keyword guard for LLM output.
A callback-driven, prefix-safe, lazy LLM stream sanitization library.
Lightweight taint tracking for LLM pipelines — label secrets at entry, block them at unsafe sinks
Security guards for LLM-powered and agentic AI applications. 35 guards, zero dependencies. Covers OWASP Top 10 for LLMs 2025 and Agentic Applications 2026.
Production-grade silent failure detection for LLM applications  hallucination alerts, PII leak detection, semantic drift, topic guard, and real-time observability
easy way to use llm watermarking techniques
easy way to use llm watermarking techniques
OWASP LLM Top 10 security scanner for AI-powered applications
LLMDet: A Large Language Models Detection Tool
LLM Detection and Response — tamper-evident audit trails for LLM traffic
Protect OpenAI and Anthropic API calls from prompt injection, jailbreaks, and data-extraction attacks.
Open-source LLM prompt injection defense — protect any AI app in 2 lines of Python
A code obfuscator to confuse LLMs
Security tool for detecting exposed LLM/AI API keys in public repositories
Predictive resource-pressure instrumentation and runtime guardrails for systems processing untrusted inputs
LLMSanitize: a package to detect contamination in LLMs
A Python library for LLM security
Lightweight validation, repair, and retry helpers for LLM outputs.
Security testing toolkit for LLM prompt injection vulnerabilities
Enterprise-grade AI security SDK with prompt injection detection for Python
A package containing storage user credential
TVP - Trust Verification Protocol. The background check of the internet.
MarkLLM: An Open-Source Toolkit for LLM Watermarking
MCP-Bastion security middleware for Anthropic Claude - prompt injection, PII, and rate-limit protection for AI agents
MCP-Bastion security middleware for AWS Bedrock - prompt injection, PII, and rate-limit protection for cloud AI agents
MCP-Bastion security middleware for Cohere - prompt injection, PII, and rate-limit protection
MCP-Bastion security middleware for CrewAI - prompt injection, PII, and rate-limit protection for AI agent crews
MCP-Bastion security middleware for Google Gemini - prompt injection, PII, and rate-limit protection
MCP-Bastion security middleware for Groq - prompt injection, PII, and rate-limit protection for ultra-fast LLM inference
MCP-Bastion security middleware integration for LangChain - prompt injection, PII, and rate-limit protection for LLM agents
MCP-Bastion security middleware for LlamaIndex - prompt injection, PII, and rate-limit protection for RAG pipelines
MCP-Bastion security middleware for Mistral AI - prompt injection, PII, and rate-limit protection
MCP-Bastion security middleware for OpenAI - prompt injection, PII, and rate-limit protection for GPT agents
MCPShield Agent - AI Agent Security Scanner for MCP servers
Healthcare-specific LLM guardrails middleware for clinical safety
Deterministic policy runtime for AI agents. Same policy file gates CI and enforces in production, with no LLM in the decision loop.
A firewall for LLMs — block prompt injection, jailbreaks, and PII exfiltration in real time.
MOSS cryptographic signing integration for Anthropic Claude
MOSS cryptographic signing integration for AutoGen
MOSS cryptographic signing integration for CrewAI
MOSS cryptographic signing integration for LangChain
MOSS cryptographic signing integration for LangGraph
MOSS cryptographic signing integration for OpenAI
Type annotations for boto3 SecurityAgent 1.44.7 service generated with mypy-boto3-builder 8.12.0
Lightweight agentic loop detector and safety monitor. No LLM required.
Open-source prompt injection firewall, hallucination blocker, and agent memory layer for any LLM app
Trust verification tool for AutoGen multi-agent conversations. Check agent safety before interaction.
Nerq trust verification for CrewAI — discover trusted agents and gate tool calls with preflight checks
Trust-gate your LangChain agents with Nerq. Preflight trust checks, agent discovery, and trust scoring for 204K+ AI agents.
Trust-check node for LangGraph agents. Verify AI agent safety before execution.
Next Policy Agent — A high-performance, secure policy engine (Python/FastAPI rewrite of OPA)
The $19/mo Permission Layer for AI Agents - Verify, Control, Audit. Block before execution.
Production-Ready Safety Layer for Agentic AI - Live observability, pre-execution safety, and audit trails
Noma guardrail middleware integration for LangChain
Agent identity primitives for Nomotic-compliant systems
OCC cryptographic proof signing for the Anthropic Python SDK
OCC cryptographic proof signing for AutoGen
OCC cryptographic proof signing for CrewAI
OCC cryptographic proof signing for Google Gemini
OCC cryptographic proof signing for LangChain
OCC cryptographic proof signing for LlamaIndex
OCC cryptographic proof signing for the OpenAI Agents SDK
Web Disable ChatGPT
Using OnlyKey as hardware SSH/GPG agent
Python SDK for Open Agent ID — register, sign, and verify AI agent identities
Python SDK for the Open Agent Trust Registry. Verifies agent identity attestations.
AgentLock authorization middleware for the OpenAI Agents SDK.
Trust & governance layer for OpenAI Agents SDK — policy enforcement, trust-gated handoffs, and Merkle audit trails
Trust-gated MCP server wrapper for OpenAI Agents SDK via Dominion Observatory
OpenAI Guardrails: A framework for building safe and reliable AI systems.
Official Python client for Open AI Guardrails policy distribution, audit evidence, and OPA control-plane APIs.
AgentLock authorization middleware for OpenClaw skill execution.
Public Preview - OpenShell + AgentMesh governance skill
Governance infrastructure that compiles and validates organization-level agent rules
OWASP harness for executable security regression testing of agentic applications and MCP-integrated systems.
OWASP Top 10 for AI Agents security assessment tools. Capabilities: full agent security scan, prompt injection detection, tool poisoning check, excessive agency, data leakage. Built by MEOK AI Labs.
The Security Observability Layer for AI Agents
Official Palveron LangChain integration — automatic policy enforcement, PII masking, and audit trails for every LLM call.
Official Palveron integration for the OpenAI Agents SDK — drop-in middleware for policy enforcement, PII masking, and audit trails.
Agent for PAM
Red-teaming large language models for train data leakage
Pangea Multipass authorization library for LangChain
Pangea Multipass authorization library for Llama Index
A wrapper around the OpenAI Python library that wraps the Responses API with Pangea AI Guard
CLI password generator and manager.
A flexible, high-performance permission storage system for Python applications
LangChain adapter that enforces PFC governance at the execution boundary
AutoGen adapter for the Phellem security framework
LangChain adapter for the Phellem security framework
Semantic Kernel adapter for the Phellem security framework
The Pilcrow LangChain adapter — deterministic compliance gate for LangChain pipelines.
SSH agent protocol with PQ keys
Post-quantum cryptographic identity and verification SDK for AI agents.