AI guardrails & agent security
3,625 agents · most starred first
AI guardrails & agent security, page 30 of 37
Static security scanner for agentic AI configuration: settings, permissions, hooks, MCP servers, agents, skills, and steering files (Checkov for AI agent stacks).
Static security scanner for AI coding agents and MCP configurations
AI Agent Runtime Security Sandbox
AI Agent security guard - policy engine for LLM tool calls
Python client for the AgentWorth governance gate (REST over the HTTP ingress)
Keyless MCP security shield: wrap any MCP server so every tools/call is screened and recorded; set the posture to enforce and a blocked call comes back as coaching your agent recovers from. Launcher for agentx-security-sdk.
Tamper-evident, third-party-verifiable receipts for AI agent / MCP tool calls
Fetch Akeyless secrets at runtime on AWS Bedrock AgentCore using cloud identity authentication
Standalone offline verifier for Open Decision Receipts (ODR) -- check schema, JCS canonical digest, Ed25519 signature, and hash-chain linkage with no Aragora install or account.
Governance patterns for autonomous AI agents in regulated banking
Standalone, dependency-light verifier for CloakLLM audit artifacts -- verify hash chains, Ed25519 signatures, key provenance, RFC 3161 timestamps, and compliance reports WITHOUT the full SDK or trusting CloakLLM's code.
AgentWatch SDK — AI agent governance, audit, and policy enforcement
Reference architecture for AI agent governance in commercial real estate operations — Autonomy Ladder™ patterns, MIT-licensed.
CrewAI tools for Signatrust — cryptographically signed, tamper-evident AI Decision Receipts.
CrewAI tools for the Trust Gate MCP server: signed receipts, a read-only-allowlist decision gate, an egress marker check and an exit drill.
LangChain tools for the Trust Gate MCP server: signed receipts, a read-only-allowlist decision gate, an egress marker check and an exit drill.
LlamaIndex tools for the Trust Gate MCP server: signed receipts, a read-only-allowlist decision gate, an egress marker check and an exit drill.
The inference boundary layer between your data and outbound AI requests
Guard CrewAI tools with EMILIA Protocol — require an offline-verifiable authorization receipt (EP-RECEIPT-v1) before an irreversible tool runs: missing -> refused, valid -> runs, replay -> refused, forged -> refused (RR-1).
Governance patterns for autonomous AI agents in regulated financial services
Finite operational ledger toolkit for AI-agent support, omissions, obligations, certificates, environment assumptions, and transitions.
LangChain and LangGraph integration for GaaS (Governance as a Service)
OpenAI Agents SDK integration for GaaS (Governance as a Service)
A governance harness for autonomous AI agents: a human proposal-gate, budget caps, an independent watchdog, and a full audit trail.
A llama based validator which checks whether a given prompt is safe/unsafe by specifying a set of policies and lists the violating policies when applicable.
Self-hosted, fail-closed approval server for AI agents (Arbiter)
Identity-based security framework. Every node is a user. Everything between is encrypted at rest.
LangChain / LangGraph integration for comply54 — African AI agent compliance enforcement
LangChain tools for Signatrust — cryptographically signed, tamper-evident AI Decision Receipts.
Tool governance and auditing framework for LangChain
A lightweight, domain-aware safety firewall for LLM applications — works with any API or local model, with dynamic policy-document support.
A compliance control layer between your app and any LLM: PII/PCI redaction, cost tracking, and privacy-safe audit logging. GCC-focused.
OWASP Top 10 for LLMs security scanner — fully offline
Auto-generate a governance ontology, Trust Profile, and signed badge for any agent, MCP server, repo, OpenAPI, Rego, or workflow. The OpenTelemetry for agent governance.
OpenAI Agents SDK tools for Signatrust — cryptographically signed, tamper-evident AI Decision Receipts.
Governance patterns for autonomous AI agents in health-insurance / payer operations (aligned to the NAIC Model Bulletin framework)
Governance patterns for autonomous AI agents in regulated payments
Governance patterns for autonomous AI agents at SEC-registered investment advisers (private capital)
Circuit breaker SDK for LLM agents.
Red-team your RAG pipeline for prompt injection and source-document leakage, in CI.
AWS IAM Identity Center OAuth 2.0 / 2.1 login implementations for Swarmauri
Salesforce OAuth 2.0 / OAuth 2.1 / OIDC 1.0 identity provider implementations for Swarmauri
DPoP proof-of-possession tooling for Swarmauri services
X.509 mutual TLS proof-of-possession verifier for Swarmauri
Universal Security Verification Framework for AI Agent Installations
Trusted dependency manager for AI agent skills
Turn the OpenAI Agents SDK human-in-the-loop tool-approval step into a verifiable, offline-checkable EMILIA authorization receipt. OpenAI pauses an agent and asks for approval; EMILIA makes that approval portable, tamper-evident evidence — a named human a
A unified contract layer for LLM agents — declarative, parameterizable contract templates enforced by deterministic checks and LLM judgment.
Reeflex adapter for Claude Code: a PreToolUse hook that gates every tool call on its impact before it runs (allow/hold/deny), fail-closed -- governance that protects you from your own agent.
Improving Zero-Shot Detection of LLM-Generated Content With Token Repetition
The runtime checkpoint for AI agents — inspect every tool call, hold the irreversible for approval, keep verifiable evidence of human oversight.
AgentRight scoring engine — pre-deployment governance assessment for AI agents
Pre-deployment governance scanning for AI agents. Scan configs, get trust scores, gate deployments.
shared rbac evaluator + cache for the 3tears platform: groups, roles, role assignments, evaluation trails
unified audit envelope, fire-and-forget publish helper, and the audit erasure rule and hub anonymization request for the 3tears platform
Inspects AI agent projects for ECZ-ID configuration and public evidence posture.
The Mandate Agent SDK — Trust Engine for the Agent Economy
Sign LangGraph tool calls and graph nodes with Vouch Credentials.
LangChain.js security integration for Weave Protocol - callbacks, tool wrappers, and retriever guards
MakerChecker connector for LangChain that wraps a StructuredTool so every invocation runs a deny-by-default governance check and records to the hash-chained audit log.
TypeScript/JavaScript SDK for the Agentum engine — agentic governance & IAM enforcement
Local-first execution governance SDK for AI agents.
TypeScript SDK for Lelu — the authorization engine for autonomous AI agents
Framework-agnostic middleware for approving or denying AI agent tool calls.
Open-source security guardrails for LLM applications — prompt injection, jailbreak, PII, hallucination & toxicity detection (with Turkish/KVKK support).
A validator for ensuring the factuality and reducing brand risk in generated content.
MCP Proxy — runtime interception for AI agent tool calls using ATR rules
Native high-performance AI agent governance proxy binary
CellFence rule plugin for AI agent change budgets.
Open-source AI agent governance for TypeScript: policy, exact one-use approvals, controlled tool execution, and verifiable receipts.
Authenticated AI agent operator API, console, and bound plan-then-execute playground.
Agent Approve plugin for OpenClaw - approve or deny AI agent tool calls from your iPhone and Apple Watch
AgentAdmit SDK: user-mediated AI agent authorization for Node.js apps
Provider-agnostic AI agent authorization with policy, approvals, rejections, durable delegation, execution leases, receipts, kill switches, and signed handoffs.
Give your AI agent a signed, verifiable identity in one line.
Scoped, signed ActionPass artifacts for runtime-governed AI agent actions.
Catch governance holes before production does. A static governance linter for AI agent workspaces.
Build an LLM agent with spending limits, a tamper-evident audit trail, PII redaction, and record/replay testing built in from the start — a governed agent in about 10 lines.
Human-in-the-loop approval gate for LLM agent tool calls. Classify tools read/write, block destructive calls behind an approver, and never persist a grant across turns.
Route LangGraph.js tool calls through toolgovern's governTool() gate before they reach ToolNode -- wraps each tool with the classifier, then re-wraps it with LangChain's own tool() factory so it slots into new ToolNode([...]) unchanged.
Kakunin compliance tools for LangChain JS/TS — verify agent certificates and scopes inside tools and chains
Mastyf AI security middleware for LangChain — intercepts and evaluates tool calls through the Mastyf policy proxy
VisIQ agent-governance harness for Node.js — action governance, retrieval governance, and audit-trail reporting, added to your existing agent framework by one visiq() call.
NoInjection enforcement adapter for LangChain structured tools.
Drop-in enforcement for AI agents — authorize every tool call through the Queldrex Trust Layer (allow / require approval / deny) before it runs, in code or as an inline MCP gateway. Fail-safe, zero-dependency.
The AI Agent Firewall for LangChain. Enforce signed authority before consequential tool calls. Free account-owned API key required.
Sphinx — local-first secret guard CLI for terminal + AI workflows
Ory Agent Security for the Cloudflare Agents SDK — per-tool authorization, tracing, and identity propagation by wrapping the AI SDK tools passed to streamText. Built on @ory/argus.
**Nyxora Policy Engine SDK** – The rigid security middleware for the Nyxora autonomous agent framework.
Duct SDK — manifest, CLI, and embed components for permissioned human and agent access to your product APIs.
Policy-as-code authorization for AI SDK tool calls, powered by [Open Policy Agent](https://www.openpolicyagent.org/).
TypeScript SDK for the Argus multi-agent security consensus oracle on Arc
eve (Vercel agent framework) tools for Trust Gate post-quantum, tamper-evident receipts.
OpenClaw integration package for Radware Agentic AI Protection
Static analyzer and GitHub Action for detecting Agentic Workflow Injection risks in AI-powered GitHub Actions workflows.
Runtime governance middleware for AI agent tool calls -- gate shell, filesystem, network, and credential access before a tool executes.
Correctover - AI Agent Runtime Assurance. CCS standard + GuardrailProvider + MCP Security Scanner. Framework-agnostic runtime security for AI agents.
Intentrax deterministic public SDK for proof verification and API-wrapper execution.
Prismor for Mastra — policy-controlled tool calls via a wrapped execute function and the Prismor eval-server
Server-first TypeScript SDK for GlobiGuard control-plane and trusted decision-engine connectivity.