MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
trust-boundary-ctfgithub.com/acseguin21/trust-boundary-ctfupdated this year
Browser-based MCP CTF — OAuth token confusion and session isolation failure (CVE-2025-49596 pattern). DevTools only.
MIT -
Credential-isolating MCP trust boundary for operating SSH hosts from Royal TS/TSX inventories
Apache-2.0 -
dlx-public-kitgithub.com/kodomonocch1/dlx-public-kitupdated this year
# DLX Public Kit DLX Phase 1 is a public evaluator kit for a thin MCP-native trust proxy. The public face is intentionally narrow: evaluators call `extract`, then `get_execution`, and judge the trust boundary by whether work ends in `succeeded` or `failed_safe`. The gateway remains behind that surface as an internal / secondary delivery layer. It
★ 1 -
MCP trust boundary — descriptor hash pinning and hash-chained evidence envelopes (optional signing) over third-party MCP tool calls.
go AGPL-3.0 -
Agentic reconciliation-break triage system — bounded LangGraph state machine, MCP trust-boundary tool split, human-approval interrupt, and a real eval harness.
★ 1 -
Autonomous AI agents inside a Qubes-isolated sandbox - tag-scoped Admin API access with dom0-mediated trust boundary.
MIT ★ 8 -
crmygithub.com/crmy-ai/crmyupdated this year
Governed context for customer-facing agents. CRMy gives AI agents trusted customer context and action boundaries in one briefing_get call. MCP-native. PostgreSQL-backed. Open source.
Apache-2.0 ★ 13 -
prioritize4megithub.com/shailipankaj/prioritize4meupdated this year
A local-first personal AI OS: a provenance-checked memory and a commitment cockpit, with a trust boundary that keeps your private data off the network.
MIT ★ 1 -
Human approval requests for agent actions: expiring, single-use decisions. Trusted executors must enforce the gate; not a same-user security boundary.
npm MIT -
life-context-vaultgithub.com/Kota-Ohno/life-context-vaultupdated this year
Give your AI the context to actually help you — without handing it your whole life. A local-first, encrypted vault that shares only reviewed Context Packs with AI clients (MCP). The trust boundary is the product.
MIT -
provekit-mcpgithub.com/Th3Circle-app/provekit-mcpupdated this year
Hardened MCP server that gives an AI agent a code security scanner, built so the tools can't be turned against the host, with a red-team suite that spawns the real server and proves the trust boundaries hold.
MIT -
Carl gives AI agents hands: a local MCP server that lets Claude, Codex and other agents use your Google account, coordinate with each other, and more — with Carl as the trust boundary.
MIT -
Find the architectural pain points in any codebase. Scores every file on layering, complexity, data access, failure handling, UI cost and trust-boundary risk against Google, Amazon and OWASP guidance. Ranked JSON for AI coding agents, an MCP server, a native viewer. Rust.
MIT -
MCP server for security-vetting third-party AI agent extensions before installation — Claude skills, ClawHub plugins, agent tool packs. v1.1: agent-config trust-boundary scanner (AGENTS.md / .cursor/rules.md / .claude/CLAUDE.md / .gemini/config / git hooks) addressing CVE-2026-26268. Outputs 0-100 risk score + BLOCK/REVIEW/CAUTION/CLEAN buckets.
pypi MIT License -
@getripple/mcpgithub.com/raushankcode/rippleupdated this year
MCP stdio server for Ripple's local authorization gate for AI coding agents.
npm MIT