JSPM

Playwright reference adapter for the SafeBrowse runtime

Cybersecify -- Security agent for AI developers. Audit OpenAPI x-agent-trust compliance, scan MCP servers, check packages, audit dependencies, verify repos, check agent threats. Works in Cursor, Claude Desktop, any MCP client. Supports the officially regi

CLI tool to manage OpenTrust AI Agent Runtime Security Platform — setup, start, stop, status, logs

Runtime security for AI Agent Skills — Scan, sandbox & enforce. Detect prompt injection, memory poisoning, supply chain attacks. 72+ patterns, 14 categories. The firewall Snyk and Cisco don't build.

Static security auditor for AI agent configs — CLAUDE.md, .mcp.json, .cursorrules. Compound threat detection, OWASP mapping, SARIF output.

Security scanner for AI agent packages — CLI + MCP server

Runtime security layer for AI agents. Enforces policy on tool calls to block prompt injection and unsafe actions.

AI Agent Security — scan every tool call for secrets, PII, destructive commands, and prompt injection. Runs locally, zero dependencies, no signup required.

Open-source AI prompt injection and data exfiltration detector. Drop-in security for LLM agents.

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

Localhost SafeBrowse daemon for browser-use agent enforcement

OpenClaw plugin for prompt injection detection. Drop-in security scanning that hooks into message, tool call, and tool result lifecycle events using heuristic pattern matching. For the full SDK with ML classification, custom rules, and audit logging, see

Algiz security plugin for OpenClaw - detect and defend against agent threats: secret masking, command guard, prompt injection scanner

Independent runtime audit and alerting for OpenClaw-like local agents.

MCP server for Vigile AI Security — query trust scores for MCP servers and agent skills from within Claude Code, Cursor, and other AI agents

OpenTrust Core — AI Agent security engine with content detection and behavior assessment

Security protocol for autonomous AI agents — contract verification, wallet monitoring, freeze protection, and x402 machine payments.

MosheSDK core engine — 7-stage pipeline, policy, analyzers, taint, chain risk

MCP server that gives AI agents 12 security tools powered by a multi-stage detection pipeline. Scans prompts, responses, SQL, file writes, and web searches for injection attacks, PII leaks, and policy violations. Session correlation engine with human-in-t

MosheSDK shared types, schemas, and validators

MosheSDK generic tool adapter — wrap any tool call through the safety pipeline

MosheSDK file-backed session, artifact, and approval store

AARI Execution Firewall plugin for OpenClaw — intercepts tool calls before execution and enforces ALLOW/WARN/BLOCK decisions

The security gateway for AI agent communication protocols (MCP, A2A). Prompt injection detection, audit logging, rate limiting, trust scoring, and policy enforcement.

TypeScript SDK for MosheSDK — runtime security for AI agents

MosheSDK Anthropic tool-use adapter — no anthropic SDK dependency required

Infinity Protocol — Cryptographic agent-human trust layer + security hardening. The loop never breaks. ♾️🐝

LangChain.js security integration for Weave Protocol - callbacks, tool wrappers, and retriever guards

Multi-tenant AI agent runtime with OS-level isolation. Sandboxed execution, encrypted credentials, private memory per tenant — one server, no Docker.

Ron Enterprise Core — 32-layer multi-tenant agentic engine

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

Thin LangGraph adapter for the OxDeAI universal execution guard

MosheSDK in-memory session, artifact, and approval store

Security scanner MCP server for AI coding agents. Prompt injection firewall, package hallucination detection (4.3M+ packages), 1700+ vulnerability rules with AST & taint analysis, LLM-powered semantic code review, auto-fix. For Claude Code, Cursor, Windsu

MosheSDK OpenAI tool-call adapter — no openai SDK dependency required

Thin OpenAI Agents SDK adapter for the OxDeAI universal execution guard

Governance guard for Claude Managed Agents. Evaluates every tool call before execution. ACCEPTED flows. RESTRICTED steers. BLOCKED stops.

Universal execution authorization guard (PEP boundary) for OxDeAI

MCPS security node for n8n -- adds cryptographic signing, tool pinning, replay protection, and audit trails to MCP workflows

SpiderRating MCP Server — security scanning and trust scores for MCP servers and agent skills

AI agent security: scan skills for 25 attack classes + runtime monitoring (EDR for AI agents). Real-time scanning, behavioral anomaly detection, session monitoring, exfiltration alerts. 1,316+ findings from 450+ audits. OWASP Agentic AI Top 10 mapped.

Prompt SecGuard — 在消息和工具数据进入 LLM 前进行安全扫描，检测提示词注入攻击

Lighthouse for AI agents. Security scanner, PII protection, prompt defense audit.

Security moat for AI agents. Runtime protection against prompt injection, tool misuse, and data exfiltration.

Evaluate AI agent plans before they execute. Returns ACCEPTED, RESTRICTED, or BLOCKED. Powered by Vouch.

VerifiedX TypeScript SDK

Security firewall for AI agents — intercept MCP tool calls, enforce policies, block attacks

MAP (Model Context Authorization Protocol) — enforcement proxy for MCP servers

the credit score for context — security scanning for packages, repos, MCP servers, skills, domains and commits

Thin AutoGen adapter for the OxDeAI universal execution guard

Security runtime for AI agents — protect against malicious tools, data theft, and prompt injection

Thin OpenClaw adapter for the OxDeAI universal execution guard

Local-first OpenClaw trust plugin for prompt-injection risk scoring, skill scanning, transcript hygiene, and posture reporting in minutes.

Zero-trust security proxy for MCP servers. Logging, rate limiting, injection detection, and policy enforcement.

Thin CrewAI adapter for the OxDeAI universal execution guard

AI agent security platform — scan, fix, monitor, and pentest MCP servers, Claude skills, Codex plugins, Cursor extensions, and 5 more platforms. 227 rules across 17 threat categories.

Agent runtime security middleware — scans tool inputs/outputs for prompt injection

Optional npm/npx wrapper for SkillGate CLI (Python core runtime)

Free, zero-friction CLI that audits MCP servers for memory poisoning vulnerabilities in <30 seconds. US-based security platform.

Security scanner MCP server for AI coding agents. Prompt injection firewall, package hallucination detection (4.3M+ packages), 1000+ vulnerability rules with AST & taint analysis, auto-fix. For Claude Code, Cursor, Windsurf, Cline, OpenClaw.

Compact instruction DSL for AI agent systems — 6-axis encoding with 4,096 instruction states in 6 tokens

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

Lightweight, zero-Python security scanner MCP server for AI coding agents. Fast install (~5s), 1700+ vulnerability rules with pure JavaScript regex engine, 4.3M+ package hallucination detection. For Claude Code, Cursor, Windsurf, Cline.

Security scanner for AI agent architectures - 220+ detection patterns and 5 runtime guard modules for prompt injection, SSRF, path traversal, credential exposure, MCP security, and OWASP ASI vulnerabilities

OCC provenance plugin for OpenClaw — cryptographic proof of every agent action.

SafeKeyLab Enterprise MCP Server - Agent Security, RAG Protection, and Compliance

Security plugin for Clawdbot. Enforces human intent boundaries inside the agent.

TypeScript client for the AumOS OWASP ASI Top 10 defensive library — agent security scanning, threat detection, and compliance reporting

Package gateway for AI agents - secure package installation with CVE scanning and threat detection

OpenClaw plugin that validates agent inputs, tool calls, and messages through the Membrane AI guard API

AI got hands. This is the leash. Policy, audit, kill switch for any AI agent with access to your accounts.

Security moat for AI agents. Runtime protection against prompt injection, tool misuse, and data exfiltration.

Security scanner for AI agent skills and prompts - detects prompt injection, jailbreaks, and data exfiltration

OpenClaw plugin for CAUL security membrane — evaluates every tool call through 4 security rings

agent-secrets-guard-server: zero-trust credential broker for AI agents

Security layer for AI agents - blocks prompt injection, credential exfiltration, and malicious commands

Official SecureClaw package has moved to @adversa/secureclaw — AI agent security hardening for OpenClaw by Adversa AI

Sentinely — Security layer for AI agents. Stop prompt injection, memory poisoning, and agent drift in 3 lines of code.

SupraWall credential vault and security layer for Paperclip AI companies. Every agent gets only the credentials it's authorized to use, for only as long as it needs them.

Drop-in behavioral telemetry for AI agents. 3-line integration, local SQLite storage, framework-agnostic.

Spectral ruleset for OpenAPI x-agent-trust compliance. Validates agent authentication security schemes per the OpenAPI Extensions Registry.

AEBA cybersecurity distribution -- Agent Event Behaviour Analytics for security teams running agent fleets. Anomaly detection + signed audit + L0-L4 trust.

AEBA for cybersecurity teams. Agent Event Behaviour Analytics -- UEBA equivalent for AI agents. Signed events, trust scoring, anomaly detection, audit chains.

AgentLair integration for Mastra — agent identity verification, behavioral trust scoring, and trust-gated tool execution.