Security & AI safety

AI guardrails & agent security

3,625 agents · most starred first

AI guardrails & agent security, page 17 of 37

@sigmashake/ssg

AI Agent Governance CLI — evaluate tool calls against rules, block dangerous operations, and surface blocked commands

@mattersec/matt

matt — AI agent telemetry for developer machines

lucifer-gate

AI agent command firewall with Telegram-based human approval

@atensec/thoth

AI agent governance SDK by Aten Security

@mnemom/agent-integrity-protocol

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

intentdna

Intent DNA — Declarative policy layer for AI agent behavior

@tankpkg/cli

Security-first package manager for AI agent skills

agent-mcp-guard

Open-source CLI scanner for risky MCP server and AI agent tool configuration.

@asqav/sdk

AI agent governance - audit trails, policy enforcement, ML-DSA cryptographic signatures

ai-aegis

AI Agent Security Monitor — intercept and approve dangerous commands in real-time

agentsec

Security auditing CLI for AI agent skills — OWASP Agentic Skills Top 10 (AST10)

@openguardrails/moltguard

AI agent security plugin for OpenClaw: prompt injection detection, PII sanitization, and monitoring dashboard

tokenauditor

Local-first MCP router integrity auditor for AI agent workflows.

agentaudit

Security scanner for AI agent packages — CLI + MCP server

@ondrej-merkun/skill-audit

Scan AI agent skills for prompt injection and malicious code

@piiiico/agent-audit

Security scanner for AI agent tooling — MCP servers, tool definitions, and agentic pipelines

agent-governance

JavaScript SDK for checking AI agent actions before they run.

optymd

Scan, score, and fix AI agent configuration files

@0xkobold/pi-whitelist

Tri-state tool permission system (allow/deny/ask) for AI agent tool invocations

nopeek

CLI for LLM agent secret safety — load env secrets without exposing values

@theajmalrazaq/agentsloopguard

Framework-agnostic safety helpers for tool-calling LLM agent loops.

mikoshi-sentinel

Deterministic action verification for LLM agent security

frenum

Deterministic, config-driven guardrails for LLM agent tool calls

request-filtering-agent

An http(s).Agent implementation that block request Private IP address.

@envsave/cli

Local secret vault for LLM-safe environments — encrypted secrets that agents can't bulk-discover

log-llm-config-staging

CLI helpers for logging hardware UUIDs and posting startup payloads to Optimus Security.

log-llm-config

CLI helpers for logging hardware UUIDs and posting startup payloads to Optimus Security.

socks

Fully featured SOCKS proxy client supporting SOCKSv4, SOCKSv4a, and SOCKSv5. Includes Bind and Associate functionality.

device-detector-js

A javascript port of Matomo device-detector

@m1a0rz/agent-identity

Agent Identity: UserPool (用户池) login, TIP token (工作负载令牌), credential hosting (凭据托管 OAuth2/API key), optional tool/skill permission control (CheckPermission) and risk approval. Integrates with Volcengine 智能体身份和权限管理平台.

@bentoguard/sdk

Bento Guard SDK - AI-powered security infrastructure for autonomous agents

javascript-obfuscator

JavaScript obfuscator

agent-security-scanner-mcp

AI agent security scanner and npm audit for MCP servers, Claude Code, Cursor, and Windsurf. Find prompt injection, hallucinated packages, secrets, unsafe tools, and vulnerable code.

@apexguard/sdk

Runtime security middleware for LLM agents — prompt injection, tool misuse, and memory poisoning defense

shieldcortex

Trustworthy memory and security for AI agents. Memory firewall for any host that writes through it; runtime tool gates on Claude Code, OpenClaw, and Hermes.

@agentlair/langchain

AgentLair adapter for LangChain.js — issue per-agent AATs, attach Bearer tokens to tool calls, emit audit envelopes, and sign outbound HTTP requests with RFC 9421 / Web Bot Auth.

@mnemopay/sdk

Governance and audit infrastructure for AI agents: charter-scoped missions, policy enforcement, human-in-the-loop approvals, tamper-evident audit chains, and EU AI Act Article 12 bundles - for money-moving and operational actions alike.

@atbash/langchain

Atbash safety guard for LangChain DynamicStructuredTool

@orangecheck/agent-langgraph

Wrap LangGraph tool nodes with OC Agent scope enforcement and emit a signed agent-action envelope per execution. Binds the action to the graph-state hash so verifiers replay the audit against the persisted graph snapshot.

ovrule-lab

TypeScript SDK for Ovrule case-file classification, guardrails, and receipt verification.

@governmyai/sdk-langchain

GovernMy.ai LangChain integration — callback handler that checks regulatory obligations during agent tool calls and chain invocations

@atbash/atbash-langchain

Atbash safety guard for LangChain DynamicStructuredTool

ssrf-req-filter

Module to prevent SSRF when making requests

@neurosec/sentry

NeuroShield Sentry — host-level agent protection daemon. Detects and blocks malicious AI agent actions at the OS level.

suprawall-langchain

Official SupraWall security integration for LangChain (TypeScript)

@silmaril-security/sdk

TypeScript SDK for Silmaril Firewall — prompt injection and jailbreak detection

protect-mcp

Local Cedar policy enforcement and signed decision receipts for AI agent tool calls.

@open-policy-agent/opa-wasm

Open Policy Agent WebAssembly SDK

@atbash/atbash-langgraph

Atbash safety guard and audit nodes for LangGraph workflows

@veridex/agent-security

Framework-agnostic agent security gateway — protect any AI agent with Veridex security packs, regardless of framework

@openai/guardrails

OpenAI Guardrails: A TypeScript framework for building safe and reliable AI systems

agents-permissions

Cross-agent permission policy specification for AI coding agents

agents-spec

Cross-agent permission policy specification for AI coding agents

agent-perms

Cross-agent permission policy specification for AI coding agents

agent-permissions-policy

Cross-agent permission policy specification for AI coding agents

@clampd/sdk

Runtime security SDK for AI agents — guard tool calls in 1 line

@agentlair/openai-agents

AgentLair adapter for the OpenAI Agents SDK — issue per-agent AATs, attach Bearer tokens to tool calls, emit audit envelopes, and sign outbound HTTP requests with RFC 9421 / Web Bot Auth.

@orangecheck/agent-openai

Wrap OpenAI function-calling (Responses API + Chat Completions tool path) with OC Agent scope enforcement and emit a signed agent-action envelope per function invocation.

@useleash/agent

Keep your AI agents on a leash. Real-time spending limits in 3 lines of code.

@tenova/swt3-ai

SWT3 AI Witness SDK: cryptographic attestation for AI inference. 103 procedures, 52 namespaces, 7 languages, 18 profiles. EU AI Act, NIST AI RMF, CMMC, SR 11-7.

tokenclaw-dev

Per-key budget caps for AI agents. Local proxy enforces hard spend limits on API keys. Scans your tools, tracks costs, blocks when over budget.

@beav3r/sdk

Beav3r client SDK for guarding autonomous agent actions.

@razroo/actionproof

Action-bound, signed proof challenges for verifying fresh autonomous agent work.

@venkat5599/x402-firewall-sdk

x402 Payment Firewall SDK - On-chain security for autonomous agent payments on Cronos

@mnemom/agent-alignment-protocol

Agent Alignment Protocol (AAP) - Verification and drift detection for AI agents

@agent_press/agentpress

agents.txt for any repo. Tell autonomous AI agents what they're allowed to do, in 60 seconds.

arcjet

Arcjet runtime security SDK — bot protection, rate limiting, prompt injection detection, PII blocking, and WAF for JavaScript and TypeScript apps

@azure-rest/ai-content-safety

ContentSafety Service

@directive-run/ai

AI guardrails and orchestration for Directive. Prompt injection, PII detection, cost tracking, multi-agent patterns.

@atbash/atbash-autogen

Atbash safety judge plugin for AutoGen-style multi-agent orchestration

meshsig

Cryptographic security layer for AI agents. Ed25519 identity, signed messages, trust scoring. Works with LangChain, CrewAI, AutoGen, OpenClaw, or any agent framework.

tibetclaw

Trust-First Agent Framework — behavioral trust, cryptographic provenance, semantic firewall. The trust kernel that agentic AI is missing.

@predicatesystems/predicate-secure

Drop-in security wrapper for AI agents - adds authorization, verification, and audit to any agent framework

@agentic-trust/core

Core SDK for agentic trust systems

@aporthq/aport-agent-guardrails-crewai

APort guardrails for CrewAI — task decorator / middleware

@verifiedstate/crewai

VerifiedState integration for CrewAI — cryptographic receipts for every agent assertion

@authorityrail/plugin-crewai

AuthorityRail governance plugin for CrewAI. Wrap any CrewAI tool or task with ARES-v1 pre-execution certification.

@scopeblind/crewai

ScopeBlind evidence wrapper for CrewAI tool calls. Automatically creates Veritas Acta receipts for every tool invocation and delegation.

@mindburn/helm-crewai-js

HELM governance adapter for CrewAI JavaScript/TypeScript — task and tool governance

@oxdeai/crewai

Thin CrewAI adapter for the OxDeAI universal execution guard

@verra/sdk

Verra AI governance SDK: detection pipeline (PII, jailbreak, prompt injection, policy violation) with LangChain, CrewAI, AutoGen, LlamaIndex, and Semantic Kernel integrations.

fullcourtdefense

Control what your AI agents do: check every tool call against your company's FullCourtDefense policy (Monitor or Enforce).

@the-bot-club/agentguard

AgentGuard TypeScript SDK with local policy contracts and compatibility HTTP/in-process adapters.

cord-engine

Secure runtime for AI agents. Policy-enforced tool execution, secret blocking, audit trails. Drop into OpenAI, Anthropic, LangChain, CrewAI. Zero dependencies.

@arikernel/adapters

Framework adapters for AriKernel — protect agents built with LangChain, CrewAI, OpenAI, and more

sentinelflow

AI agent governance platform — static scanning + runtime interception for Claude Code, Cursor, GitHub Copilot, Codex CLI, LangChain, CrewAI, and Kiro. Blocks dangerous tool calls in real-time.

dingdawg-loop

Governance receipts for AI agents — DingDawg Loop Protocol (DDLP) for safe scheduled agents with governance gates, compliant with EU AI Act Art. 12, Colorado AI Act (SB 24-205), and NIST AI RMF. Every loop execution is verified, receipted, and fail-closed

@axonflow/sdk

AxonFlow SDK - Add invisible AI governance to your applications in 3 lines of code

agent-governance-check

Five governance questions for your AI agent system. Identity, restraint, accountability, memory, charter. Zero dependencies.

miii-security

Security-focused `SKILL.md` packs for reviewing and hardening LLM systems.

@connexum/ai-governance

Enterprise AI agent governance framework. Hook-based enforcement, compliance packs (SOC 2, HIPAA, GDPR, PCI DSS, DORA, EU AI Act, ISO 27001), audit trails, multi-LLM adapter architecture (Claude production today; additional providers on roadmap).

@damitha-perera/agentgate

Authorization middleware SDK for AI agents — powered by AgentGate

buzur

AI prompt injection defense scanner for agents. 25 phases covering HTML obfuscation, encoding attacks, semantic similarity, behavioral anomaly detection, multi-step attack chains, supply chain worm detection, and more.

@vorim/agent-audit

Free AI agent code hygiene checker. Scans for hardcoded LLM keys, shared credentials, and long-lived agent permissions in TypeScript, JavaScript, and Python projects.

@varpulis/agent-runtime

Real-time behavioral guardrails for AI agents. Detect retry storms, circular reasoning, budget overruns, and failure spirals — as they happen.

@oxdeai/guard

Universal execution authorization guard (PEP boundary) for OxDeAI

curiosity-cat

AI agent safety framework — standing orders, danger maps, and stories that teach cats to land on their feet

agent-airlock

Human approval gate for AI agent actions. Drop-in protection against prompt injection for any API key.

@mindburn/helm-ai-kernel

TypeScript SDK for HELM — fail-closed tool calling for AI agents

@gns-aip/sdk

AI Agent Identity Protocol — Proof-of-Jurisdiction for AI agents on the GNS Protocol

More in Security & AI safety