Security & AI safety

AI guardrails & agent security

3,625 agents · most starred first

AI guardrails & agent security, page 18 of 37

@agentbouncr/core

Agent Governance Framework — Core Engine (Permission Layer, Policy Engine, Audit Trail)

@gridseal/core

Tamper-evident, cryptographically verifiable audit trail for AI decisions

@futurespeak-ai/claw-framework

Constitutional AI Governance Framework — Asimov's cLaws with HMAC-SHA256 integrity verification, memory watchdog, safe mode, and Ed25519 cross-agent attestation

delegate-os

Scoped delegation protocol for the agentic web — DCTs, attestations, and MCP middleware

trustloop

AI governance SDK — intercept, audit, and control AI agent tool calls. Add human-in-the-loop approval, kill switches, and a full audit trail to any AI agent.

@metatransformer/agentguard

Security scanner and trust layer for AI agent skills/plugins — scan URLs, files, and directories for credential exfiltration, injection, and other threats

vorloop-agent-safety

Safety infrastructure for autonomous AI agents

@inkog-io/mcp

Security co-pilot for AI agents. Scan for vulnerabilities, verify governance, audit MCP servers, and generate compliance reports — all from Claude, Cursor, or any MCP client.

@rakesh90466/aip-sdk

Zero‑dependency Agent Identity Protocol (AIP) SDK – Ed25519‑based token signing & delegation.

@interven/gateway

Interven AI Agent Firewall — drop-in gateway proxy. Scan every tool call through policy + risk scoring. One command, zero infrastructure.

@rubric-protocol/sdk

Rubric Protocol SDK - post-quantum AI attestation for EU AI Act compliance. Patent Pending.

firmis-cli

The security layer for AI agents — platform-agnostic threat detection with 300+ rules, runtime blocking, and remediation guidance. Continuous protection.

@guard0/g0

Background check for AI agents — discover, assess, and test before you ship

dingdawg-governance

Governance receipts for AI agents — universal, MCP-native, fail-closed governance layer compliant with EU AI Act Art. 12, Colorado AI Act (SB 24-205), and NIST AI RMF. LNN interpretability, IPFS audit proofs, and rollback for any agent in any framework.

@antenehtessema/dasc-core

Deterministic Agentic Safety Controller (DASC) for Node.js

@deepsweepai/mcp-firewall

Enterprise-grade MCP firewall for conscious AI stewardship - Zero-trust security layer for Model Context Protocol servers

a1-ai

TypeScript/Node.js client for A1: cryptographic identity and authorization for AI agents

arkna

Official Node.js SDK for Arkna: AI agent evidence

@vorim/sdk

Official TypeScript SDK for Vorim AI — AI Agent Identity, Permissions & Audit

@interven/sdk-js

Interven JS/TS SDK — Approvals + Audit for AI Agents. Scan tool calls through policy + risk scoring. Block, redact PII, or route to human approval via Slack.

@phoenixaihub/agent-contracts

Design by Contract for AI Agent Steps — runtime preconditions, postconditions, and invariants for agent execution

@audit0/sdk

Zero-config compliance logging for AI agents. Real-time audit feeds, risk scoring, and human-readable summaries.

@atpdevelopment/openclaw-atp

ATP (Agent Trust Protocol) security integration for OpenClaw agents

agentfortress

Runtime protection and security monitoring for AI agents — The CrowdStrike for AI Agents

@agentsentinel/sdk

Safety controls for AI agents — budgets, approvals, rate limits, audit logs, and DLP

@arikernel/middleware

Drop-in middleware wrappers for securing AI agents with Ari Kernel

@agentsecretstore/sdk

Official SDK for Agent Secret Store — the secure vault built for AI agents. Store and retrieve credentials with scoped tokens, envelope encryption, and audit trails.

aip-protocol

The only thing that makes AI agents production-ready. Cryptographic identity, boundary enforcement, and kill switch for autonomous agents.

@meshsig/sdk

Cryptographic security layer for AI agents. Verify instruction origin before execution.

cyphrex

Security and identity infrastructure for AI agents

taintctl

Content-aware provenance layer for Claude Agent SDK. Detects dangerous content at every tool I/O boundary, propagates classification across sub-agent dispatch, and enforces fail-closed policy.

@gethelio/proxy

Open-source governance for MCP agents: useful autonomy without unlimited authority

@aporthq/aport-agent-guardrails-langchain

APort guardrails for LangChain/LangGraph — callback handler

@atbash/langgraph

Atbash safety guard and audit nodes for LangGraph workflows

@oxdeai/langgraph

Thin LangGraph adapter for the OxDeAI universal execution guard

@graphos-io/sdk

GraphOS SDK — wrap your LangGraph app with policy enforcement and live telemetry

@verifiedstate/langgraph

VerifiedState integration for LangGraph — cryptographic receipts for every agent assertion

occ-langgraph

OCC cryptographic proof signing for LangGraph node executions

@authorityrail/plugin-langgraph

AuthorityRail governance plugin for LangGraph. One import = ARES-v1 Certified Action Records on every node.

psy-core

Tamper-evident audit log for AI agent memory operations. Adapters for Anthropic Memory, Letta, Mastra, Mem0, LangChain, LangGraph, GBrain, and Hermes Agent (via psy-core-hermes).

@oconnectortechnology/abs-langgraph

ABS CORE Governance Adapter for LangGraph/LangChain

@letsping/sdk

Agent trust layer: behavioral firewall, HITL, and Cryo-Sleep state for AI agents. Works with LangGraph, Vercel AI SDK, and custom runners.

@authensor/langchain

Authensor guardrail adapter for LangChain/LangGraph

AgentEnc

飞桨模型加密库

AgentLogin

A project for Third party login

LLM-safeguard

Universal middleware for safeguarding AI‑generated content

SecureAgent

Agentic Identity Framework abstracting Keycloak/OIDC complexity for AI Agents

accuknox-llm-defense

Python SDK for AccuKnox LLM Defence API

acp-crewai

Agentic Control Plane governance for CrewAI agents. Wrap any tool with @governed; ACP decides allow/deny/redact.

acp-langchain

Agentic Control Plane policy & audit for LangChain / LangGraph / Deep Agents. One ACPMiddleware() on create_agent; ACP decides allow/deny/redact on every tool call, subagents included.

actiongate-langchain

Gate LangChain tool execution through ActionGate primitives before execution.

actiongate-openai

Gate OpenAI function/tool calls through ActionGate primitives before execution.

aegis-agent

Human-in-the-loop control layer for AI agents

afterdark-llm-firewall

Production-ready security layer for AI applications

agent-action-guard

Runtime classifier for screening AI agent actions as safe, harmful, or unethical.

agent-action-policy

Declarative action policies for AI agents — approve, deny, or escalate any tool call before execution

agent-aegis

Auto-instrument AI agent frameworks with runtime security. One line to govern LangChain, CrewAI, OpenAI, Anthropic — injection blocking, PII masking, action policy, audit trail. Zero code changes.

agent-airlock

A type-checker and contract layer for AI agent tool calls: in-process, Pydantic-only. Strict argument validation, ghost-argument stripping, and self-healing retries for MCP servers and agent frameworks.

agent-audit

Static security analyzer for AI agents — prompt injection, tool input validation, MCP config auditing, secret detection. 66 rules mapped to OWASP Agentic Top 10.

agent-audit-logger-mcp

Hash-chained HMAC-signed audit log MCP for A2A (agent-to-agent) calls. Every tool-call, agent-handoff, decision gets a tamper-evident signed record. EU AI Act Art 12 automatic logs, DORA Art 17 ICT incident logs, ISO 42001 clause 9 monitoring — auditor-ready end-of-day attestations. By MEOK AI Labs.

agent-audit-trail

Tamper-proof compliance audit trail for AI agents. EU AI Act Article 12 + SOC 2 ready.

agent-audit-trail-mcp

Immutable audit logging for AI agents — hash-chained event log, integrity verification, EU AI Act compliance

agent-ca

A Zero Trust Cryptographic Certificate Authority SDK for Autonomous AI Agents.

agent-canary

Scan your agentic codebase for unguarded tool calls with real-world side effects

agent-challenge

LLM-solvable challenge-response authentication for AI agent APIs

agent-context-guard

Integrity verification and access control for AI agent context files

agent-control-plane

Layer 3: The Framework - A governance and management layer for autonomous AI agents

agent-control-sdk

Python SDK for Agent Control - protect your AI agents with controls

agent-control-server

Server for Agent Control - manage and evaluate controls for AI agents

agent-did-sdk

Python SDK for the Agent-DID Specification (RFC-001) — create, sign, resolve, verify and revoke Agent-DIDs

agent-dna

Security, Governance and Audit framework for AI Agents

agent-execution-guard

A lightweight execution guard for AI agents. Decide ALLOW / HOLD / DENY before your agent performs real actions.

agent-failsafe

FailSafe governance adapter for Microsoft Agent Governance Toolkit

agent-gate-sec

AgentGate - AI Agent 数据采集与安全分析中间层服务

agent-gatekeeper

Identity and permission layer for AI agents — Auth0 for agents

agent-gov

AI agent governance and compliance engine with policy enforcement

agent-governance

Agent governance tooling

agent-governance-cerone

A lean governance system for AI agents.

agent-governance-cerone-sdk

A lean governance system for AI agents.

agent-guard-core

Python library to secure AI agents with environment credentials from multiple secret providers and more

agent-guard-plugins

Drop-in prompt-injection guards for Claude, OpenAI Codex, Hermes, and OpenCLAW agents. Wraps the agent-guard-modernbert-base and agent-guard-deberta-pi-base classifiers on Hugging Face.

agent-guardrail

Action-level governance for AI agents -- control what they DO, not what they SAY

agent-guardrails

A package for LLM guardrails

agent-guardrails-zhuyt

🛡️ Lightweight Agent Guardrails - Make small local LLMs reliable for production agent workflows

agent-iam-ratchet

Sensitivity Ratchet SDK — irreversible permission narrowing for AI agents

agent-id-cli

AgentID — CLI for managing AI agent identities

agent-id-client-sdk

AgentID — Client SDK for AI agents

agent-id-service-sdk

AgentID — Service-side SDK for verifying AI agent JWTs and reporting activity

agent-identity

Agent identity management, trust scoring, and certificate management

agent-identity-bridge

Portable identity for AI agents across MCP, A2A, ANP, and AG-UI protocols

agent-identity-cli

Agent Identity Toolkit CLI for managing AI Agent identities

agent-identity-protocol

Verifiable cryptographic identity and delegation for AI agents across MCP and A2A

agent-identity-python-sdk

Python SDK for using Agent Identity Service

agent-identity-registry

Python SDK + CLI for the Agent Identity Registry (AIR) — agentidentityregistry.org

agent-ids

Agent identity primitives for AGTP-compliant systems

agent-immune

Adaptive threat intelligence for AI agent security — semantic memory, multi-turn escalation, output scanning, rate limiting, and prompt hardening.

agent-integrity-proto

Agent Integrity Protocol — real-time thinking block analysis for AI agent alignment

agent-intent-guard

MCP tool-call guardrails with static and semantic policy checks

agent-memory-guard

Runtime defense layer that protects AI agent memory from poisoning, tool abuse, privilege escalation, and excessive autonomy (OWASP).

agent-moss

Multi-layer security analysis engine for AI agents

More in Security & AI safety