tryb

Evidence-first beta

MCP Security Observatory

The complete latest official Registry sweep, separated from static and execution scan coverage. Absence of findings is never presented as a safety claim.

Security Index

Pending

0 comparable scans

Monitored

0

Public server identities

High risk

0

Latest completed scan

Open findings

0

Public, non-embargoed

Disclosures

0

Responsible disclosure records

30229 canonical servers

Page 127 of 1512

rankedtoday-mcp

This MCP server provides League of Legends math and cognitive-readiness tools for AI agents, including MMR estimation, climb projection, KDA calculation, and sleep/reaction time research. It requires no API key or network calls, as every tool is a pure function with visible arithmetic.

vunversioned · static queued · execution queued

Unscanned

reader

MCP server that lets Claude send documents, drills, and choices to an e-reader (Kindle, Kobo, etc.) in real time via a Cloudflare Worker, supporting anonymous and OAuth-authenticated access.

vunversioned · static queued · execution queued

Unscanned

rokadoc MCP Server

Provides a Model Context Protocol interface to NTT Docomo's rokadoc RAG service, enabling document conversion and semantic search from MCP-compatible clients.

vunversioned · static queued · execution queued

Unscanned

santismm-knowledge

Provides a read-only MCP interface to the Santismm Knowledge Platform, offering tools to retrieve curated content on engineering, AI patterns, architectures, governance, and agent taxonomy with structured output in English, Spanish, and Portuguese.

vunversioned · static queued · execution queued

Unscanned

snow-mcp-server

Secure MCP runtime on Google Cloud Run enabling ServiceNow AI Agents to access another ServiceNow instance as the signed-in user via delegated OAuth, with read-only tools for work items and pending approvals.

vunversioned · static queued · execution queued

Unscanned

soft1-mcp

Enables read-only analysis of Soft1 ERP databases through guarded T-SQL queries, schema discovery, knowledge search, and CSV/XLSX exports.

vunversioned · static queued · execution queued

Unscanned

sol-indesign-mcp

Enables OpenAI Codex to inspect and modify Adobe InDesign documents through a local, authenticated MCP bridge and UXP panel. It provides typed, allowlisted operations with workspace confinement, approvals, and one-step Undo.

vunversioned · static queued · execution queued

Unscanned

speclaw

Local MCP server that gives AI agents a project constitution (LAWS.md), a code graph (Compass), and a spec-driven workflow — all 100% local with no LLM or cloud.

vunversioned · static queued · execution queued

Unscanned

ssap-mcp

An MCP server that automates the SSAP2010 slope stability analysis program, allowing AI assistants to create models, select calculation methods, run verifications, and read the factor of safety results.

vunversioned · static queued · execution queued

Unscanned

stagehand-repl-mcp

Provides LLM agents with a persistent, scriptable browser using Stagehand and Playwright, exposing a single stagehand_eval tool for running JavaScript snippets and AI-driven browser automation.

vunversioned · static queued · execution queued

Unscanned

Stalled Fulfillment Recovery MCP Server

An MCP server that enables operations teams to diagnose and recover stalled fulfillment orders via tools like retry_fulfillment and escalate_order, with safety boundaries limiting writes to retry and audit operations.

vunversioned · static queued · execution queued

Unscanned

stateless-mcp-incident-lab-typescript-raw

A raw TypeScript MCP server simulating incident lifecycles with tools, resources, prompts, and signed multi-round remediation approval. Enables CLI-based interaction, protocol negotiation, and health/security testing without external MCP SDK dependencies.

vunversioned · static queued · execution queued

Unscanned

SubMCP

MCP server that delegates bounded subtasks to NVIDIA NIM sub-agents, enabling context offload and parallel fan-out for coding assistants like Claude Code and Cursor.

vunversioned · static queued · execution queued

Unscanned

SummitFlow MCP

Grant management tools for AI agents: plan backward from deadlines, count business days around US federal holidays, build budgets with indirect costs, and value volunteer and in-kind contributions. No API key or network required—all pure functions.

vunversioned · static queued · execution queued

Unscanned

Todo MCP

A demo MCP server that exposes a shared todo list as tools over Streamable HTTP, enabling agents to list, add, complete, and delete todos alongside a Next.js web UI.

vunversioned · static queued · execution queued

Unscanned

Tranna MCP

A production-oriented MCP server foundation for connecting Claude Desktop to social-media and marketing platforms, featuring a reusable connector framework and an optional Meta connector using official APIs.

vunversioned · static queued · execution queued

Unscanned

twocents

A remote MCP server that lets AI agents share HTML pages with real people for visual review and collect their pinned notes as actionable feedback.

vunversioned · static queued · execution queued

Unscanned

type-atlas

Enables coding agents to navigate TypeScript, Markdown, and JSON with editor-grade features such as definitions, references, diagnostics, and code actions via the Model Context Protocol, using the local TypeScript project for each file.

vunversioned · static queued · execution queued

Unscanned

umami-recipes-mcp

MCP server for Umami Recipes that enables listing recipe books, searching and reading recipes, and creating or editing recipes with atomic tag bookkeeping. Supports remote Streamable HTTP on Cloudflare Workers and local stdio modes.

vunversioned · static queued · execution queued

Unscanned

vault-server-MCP

Remote MCP server for Obsidian vault access, giving Claude read/search/archive access to markdown notes via OAuth 2.1 + PKCE auth.

vunversioned · static queued · execution queued

Unscanned

Catalog coverage

Every version record returned by the last completed official Registry cursor sweep.

Static coverage

Metadata and artifact checks completed for the current content-bound revision.

Execution coverage

Sandbox behavior checks; unavailable until a conforming external worker completes them.