- s-gw @sgateway/s-gwKeep credentials out of AI coding agents with local approvals, scoped injection, and audit trails.6% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.15 GitHub stars.
- Ntobjmanager @lupingqaq/ntobjmanager-mcpStateful Windows RPC attack-surface research for AI agents: 22 tools on NtObjectManager.Not assessed yet.Runs on your machine, installed from a package registry.14 GitHub stars.
- Mac Shortcuts — read your Mac from any agent @dominic-dk/mac-shortcuts-mcpRead your Mac's Calendar, Reminders, Notes, Contacts and more via Shortcuts. Read-only, local.7% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.14 GitHub stars.
- Munim Computer Use @munimtechnologies/munim-computer-useAccessibility-first desktop control for any agent: background input, agent pointer, zoom, Chrome.6% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.14 GitHub stars.
- Orbit @constellation-works/orbitOrbit's local-first task and workflow registry for coding agents.4% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.14 GitHub stars.
- WinDbg @glslang/windbg-mcpWinDbg/DbgEng over MCP: crash dumps, live user & kernel, driver IOCTLs, and TTD.8% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from MCPB.14 GitHub stars.
- Unterm @zhitongblog/untermThe terminal AI agents can drive: spawn panes, run commands, read screens, scrolling screenshots.6% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from a package registry.14 GitHub stars.
- Android @us-all/androidAndroid MCP — ADB-based device, apps, UI, logcat, emulator, files, system, debug7% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.14 GitHub stars.
- Memorydetective @carloshpdoc/memorydetectiveiOS leak/perf debugging via MCP: memgraph cycles, .trace analysis, SourceKit-LSP bridging.3% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.14 GitHub stars.
- ProPresenter @alxpark/propresenter-mcpProPresenter 7 MCP server5% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.14 GitHub stars.
- Automox MCP Server @automox/automox-mcpOfficial MCP server for Automox. Manage devices, patches, and policies in natural language.3% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from PyPI.14 GitHub stars.
- Pointframe MCP Server @dimitar-radenkov/pointframe-mcpWindows screenshots, OCR, and redacted screen recording for AI agents. No desktop app needed.4% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from MCPB.13 GitHub stars.
- Lumae @matlegault/lumaeRecord, cut, zoom, caption and export software demos from an agent. Requires Lumae for macOS.5% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from MCPB.13 GitHub stars.
- Async Parallel Antigravity for Codex & Claude Code @varadfromeast/codex-agy-bridgeRun parallel, resumable, human-operable Antigravity CLI sessions from any MCP agent harness7% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from PyPI.13 GitHub stars.
- Clawtouch @tinqiao-oss/clawtouch-mcpUSB-HID keyboard/mouse (Pico 2, open firmware) as MCP tools; --mock runs with no hardware5% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from PyPI.13 GitHub stars.
- Glovebox @segentic-lab/glovebox-mcpSandboxed computer-use for AI agents: drive real browser + desktop apps in nested X11 windows6% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from a package registry.13 GitHub stars.
- Infrabroker @luisgf/infrabrokerSSH & Kubernetes access broker for AI agents; the model never touches a credential.7% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from Docker.13 GitHub stars.
- WinBridgeAgent @codyard/winbridgeagentOpen-source Windows MCP server for AI assistants. File, clipboard, screenshot, process & power mgmt.14% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from a package registry.13 GitHub stars.
- Execkit @blinkingbit-oss/execkitStateful, structured, safe shell sessions for AI agents over local, SSH, and Docker.4% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from PyPI.12 GitHub stars.
- Osascript @m0rvayne/mcp-osascriptmacOS automation: 17 typed tools for windows, menus, keyboard, clipboard, tabs, screenshots4% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.12 GitHub stars.
- Aletheia Runtime Safety & Scope Creep Filter @vikasny30/aletheia-mcpDeterministic pre-execution filter that blocks known scope-creep and prompt-injection tool calls5% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.12 GitHub stars.
- Tailscale @tailscale-mcp/tailscale-mcpDrive a Tailscale node and its tailnet through the CLI and the control-plane API4% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.12 GitHub stars.
- Editmamei @editmamei/editmameiPhotoshop MCP server: natural-language AI photo editing in desktop Photoshop. AI orchestration.5% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.12 GitHub stars.
- Smart Terminal @pungggi/smart-terminalMCP PTY server providing AI agents with real interactive terminal access5% chance this is malicious. Assessed by AI, which can make mistakes.Runs on your machine, installed from npm.12 GitHub stars.
