AI guardrails & agent security
3,625 agents · most starred first
AI guardrails & agent security, page 25 of 37
Quantum-resistant credential wallet for AI agents. ML-KEM-768 key encapsulation, AES-256-GCM at rest, ML-DSA signed access audit log. LangChain/AutoGen/CrewAI integrations.
PQC-signed RAG pipeline chunks. Sign document chunks with ML-DSA at ingestion, verify at retrieval. Prevents vector database poisoning.
Offensive security testing for AI agents
Prompt Defender. A package to help you defend against prompt injection attacks.
Production-ready LLM security firewall powered by Groq
ProofGate Action Provider for Coinbase AgentKit - Blockchain guardrails for AI agents
Evals and observability for RAG: prompt-injection screening on stored documents
purifyllm
A security boundary for LLM agents — policy, audit, and rollback for every tool call
Helmet.js for AI Agents — Lightweight security middleware for production AI agents
Prompt injection detection for Agents using ZEDD (Zero-Shot Embedding Drift Detection)
EYDII Verify tools and middleware for Pydantic AI — verify every agent action before execution
Forge Verify tools and middleware for Pydantic AI — verify every agent action before execution
Production-ready guardrails for Pydantic AI with native integration patterns
Guardrail capabilities for Pydantic AI — cost tracking, tool permissions, input/output guards
Deny-first tool authorization for pydantic-ai agents — by AgentsID
Benign slopsquatting research package
Governance middleware for PydanticAI — semantic policy enforcement, trust scoring, and audit trails for agent tool execution
Easy to use Python library for detecting AI-generated text
Python implementation of Vault Agent with client-side caching and automatic authentication
Qortara Governance — LangChain sidecar SDK. Deep tool-dispatch interception + Portable Trust Attestations + policy enforcement companion to LangSmith.
Quilr Guardrails Integration for LiteLLM
Behavioral telemetry and policy enforcement for AI agents
Middleware that prevents sensitive information leakage in RAG pipelines by analyzing the context assembly graph
Client-side retrieval firewall for RAG systems — blocks prompt injection and secret leaks, re-ranks stale or untrusted content, and keeps all data inside your environment.
Enterprise-grade data poisoning detection & alerting for RAG systems
Scan RAG documents for hidden prompt injections, invisible text attacks, and data exfiltration payloads before they enter your vector database.
A runtime governance, security and execution control layer for Retrieval-Augmented Generation (RAG) systems.
Guardrails module for Ragbits components
Audit trail middleware for RAG pipelines in regulated industries
Efficient defense against knowledge corruption attacks on RAG systems
Security middleware for RAG pipelines — detect adversarial hallucination attacks before they reach your LLM.
CLI tooling for Rare agent identity flows
A library that connects to ams using argo-ams-library and syncs with MITREid, SimpleSAMLphp and Keycloak
Adversarial testing for LLM applications. Pip install. Async-first. Reproducible.
Runtime interception SDK for AI agent governance
Generate agentic workflows visualizations & threat modelling reports for your agentic projects.
Comprehensive security toolkit for LLM applications - https://resk.fr
Agent authentication and orchestrator client utilities for secure multi-agent communication
Rubrik Agent Cloud policy enforcement plugin for AI agent frameworks
Coming soon
A safer way to use GPT chat models
AI agent execution guard with durable claims, request-id deduplication, unresolved-outcome handling, and receipts.
World's Most Comprehensive Secure Agentic AI Framework - 12 Security Foundations, Continual Learning, Multi-LLM Support
Universal Guardrails SDK for Large Language Models
A lightweight Python guardrail SDK for content safety
Using SafeWISE as hardware SSH/GPG agent
Using hardware wallets as SSH/GPG agent
Microsoft AutoGen Python tool that gates KeeperHub workflow execution through SBO3L's policy boundary. Composable with sbo3l-sdk.
CrewAI tool wrapping SBO3L — gate every Crew action through SBO3L's policy boundary.
CrewAI Python tool that gates KeeperHub workflow execution through SBO3L's policy boundary. Composable with sbo3l-sdk.
LangChain Python tool wrapping SBO3L — gate every agent payment intent through SBO3L's policy boundary.
LangChain Python tool that gates KeeperHub workflow execution through SBO3L's policy boundary. Composable with sbo3l-sdk.
LangGraph adapter for SBO3L — drop a PolicyGuardNode between your plan and execute nodes to gate every action through SBO3L's policy boundary.
LlamaIndex tool wrapping SBO3L — gate every agent payment intent through SBO3L's policy boundary.
Ed25519 signed receipts for LlamaIndex agent tool calls
Ed25519 signed receipts + Cedar policy enforcement for Pydantic AI agents
Ed25519 signed decision receipts for kyegomez/swarms multi-agent systems. Wrap any tool in a Swarms agent with tamper-evident, offline-verifiable receipts in the Veritas Acta format (IETF draft-farley-acta-signed-receipts).
Drop-in LLM secret scrubber to prevent API key and credential leaks
An agent for handling secrets in config
Tools for extract sensitive configuration out of your project
Typed secret lookup from environment variables and HashiCorp Vault
Reserved package name for Secure Agentics
AI Security Operations Platform — Python SDK & Security Gateway
Python SDK for Security Agent specifications (models and JSON Schemas)
Local-first DLP engine and security layer for AI coding agents. Secures Copilot, Claude Code, Cursor, Windsurf, and any MCP-compatible agent.
Python SDK for Self Agent ID — proof-of-human verification for AI agents
SENTINEL — AI Security Platform. 49 Rust Engines + Micro-Model Swarm. Defense, Offense, Framework.
Generates an AWS IAM session policy given a description of the tasks the user needs to perform.
A small example package
KYA (Know Your Agent) identity verification for HuggingFace smolagents
Governance gate for LangChain agents. Powered by Sentinel AI — pauses risky actions for human approval, logs every decision, generates compliance reports.
Rules-as-Markdown AI agent governance validator. 35 compliance checks, A2A v1.0 (Linux Foundation) compatible, soul.py powered.
LangGraph node bindings for SoulPolicy drift / guardrail primitives
Runtime security for AI agents — detects jailbreak, prompt injection, and tool misuse from inside the agent.
This is used as a placeholder to help prevent dependency confusion attacks.
This is used as a placeholder to help prevent dependency confusion attacks.
Drop-in middleware that gives AI agents hard spending limits, token budgets, circuit breakers, and real-time kill switches.
A Python keyring backend unlocked by ssh-agent signatures
LLM sanitization SDK — DOMPurify, but for LLM context windows.
AgentMesh adapter for structural authorization gates — consumes external policy decisions, trust grades, and delegation scope chains as AGT trust signals
Drop Stryda governance into a LangChain agent — wrap a BaseTool, or attach a callback to an AgentExecutor.
Stryda governance as a LangGraph node — plug it between the agent LLM and the tool executor.
Runtime guardrail for tool/function calls: policy, optional approval, and a verifiable ledger.
Adversarial resilience testing for multi-agent ecosystems
SWARM: System-Wide Assessment of Risk in Multi-agent systems - A Distributional AGI Safety framework
Swarm Shield - TGSC
Apple OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
Azure Active Directory OAuth 2.0 / 2.1 / OIDC 1.0 login implementations for Swarmauri deployments
AWS Cognito OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
Facebook OAuth 2.0 / OAuth 2.1 / OIDC 1.0 login implementations for Swarmauri
Gitea OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
GitHub OAuth 2.0 / OAuth 2.1 identity provider implementations for Swarmauri
GitLab OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
Google OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
Keycloak OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
Okta OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
ACME certificate service for Swarmauri
CFSSL-backed certificate service for Swarmauri
OCSP certificate verification service for Swarmauri.