Security scanner for MCP servers. Grades authentication, permissions, input validation, and tool safety. The Lighthouse of agent security.

Safe builders for Trusted Types values

Strict TypeScript type definitions for the WebMCP core API (document.modelContext)

Bigmap for Keyv

Unified LLM API with automatic model discovery and provider configuration

Shared OpenFeature JS components (server and web)

Strict WebMCP core runtime polyfill for document.modelContext (with deprecated navigator.modelContext alias)

A type-safe marriage of `EventTarget` and `EventEmitter`.

Agentic loop system with tool execution for LLMs

A safe migration checker for Prisma.

Small module that makes sure your catch, caught an actual error and not a programming mistake or assertion

Openapi for zodios

Model Context Protocol Server for Rebase — exposes schema, DB, document, and user management tools to AI assistants.

A deep deletion module for node (like `rm -rf`)

JavaScript build tool, similar to Make or Rake

A unified, high-performance i18next CLI.

AI safety guardrails for OpenClaw: CLAW protocol validation, data leak prevention, and tool safety for personal AI agents

AI red-teaming for the AI agents & LLM apps you own: stress-test them with adversarial models to find security failures (prompt injection, tool misuse / excessive agency, data leakage, jailbreaks, denial-of-wallet), auto-patch (blue team), retest, and get

Unified LLM API with automatic model discovery and provider configuration

Vendo's policy core at one choke point: guard.bind, approvals, grants, audit, Vendo Auto judge, breakers.