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.
-
@getripple/mcpgithub.com/raushankcode/rippleupdated this year
MCP stdio server for Ripple's local authorization gate for AI coding agents.
npm MIT -
Human Design API. Bodygraph in one call: energy type, strategy, inner authority, profile, definition, incarnation cross, nine centers, channels, gates.
MIT ★ 1 -
Evidence-based merge gate for AI-authored PRs. Exit codes lie; prove green with output.
npm Apache-2.0 -
Policy-aware MCP server for Keel: expose on-chain agent actions to a model, gated by the authorization policy before anything executes.
npm Apache-2.0 -
Agentic QA framework: Claude authors, generates, triages & self-heals Playwright tests. MCP server + CLI, gated in GitHub Actions.
npm MIT ★ 5 -
Local MCP/AI attack-surface audit and evidence gate for authorized testing
NOASSERTION ★ 269 -
mcp-for-metabasegithub.com/dominikpollok/mcp-for-metabaseupdated this year
Safety-gated MCP server for Metabase API access and dashboard authoring
pypi GPL-3.0-or-later -
Open-source AI agent security gate: deterministic ALLOW/BLOCK/ESCALATE authorization before every tool call, plain-English policy compilation, tamper-evident audit ledger. MCP security gateway + Claude Code/Cursor guardrails. Rust, Apache-2.0.
Apache-2.0 -
Commerce Gate MCP provides eleven commerce-action preflights, D365 authorization, and policy evidence. Its eight default tools make no marketplace or ERP writes. Optional CommerceAgent run tools are disabled by default; write proposals require a matching
npm Apache-2.0 -
Local-first, authorization-gated attack-surface intelligence with live, parallel discovery, traceable assets, and read-only MCP access.
Apache-2.0 ★ 1 -
PRAMAN (प्रमाण) — evidence-driven verification toolkit for AI coding agents. Authored browser/command checks, strict proof gates, CLI + stdio MCP.
MIT -
Sekudo, a scope-gated DAST exploitation MCP server. Hunts hard inside authorized targets and proves findings with benign proof-of-concept.
npm MIT -
MCP server where every AI agent tool call is authenticated and authorized by a Keycloak-issued JWT: role-gated tools, clear Forbidden errors, audit log
MIT -
Microsoft 365 Outlook Mail via delegated Microsoft Graph: read messages, folders and attachments, and author/manage drafts (send is a later gated phase).
npm MIT -
Reproducible MCP tool-poisoning (line-jumping) demo, contained by an independent plan-vs-authorize gate. Own testbed, mock server, fake secret — attacks nobody.
MIT -
Verify agent-authored change claims at exact Git commits with an evidence-first PR gate and local MCP.
npm MIT -
Agentic book-writing workflow: grounds drafting in an author's own past prose pulled from Google Drive, then gates every chapter on canon, continuity and style fidelity. TypeScript, MCP, OpenSpec.
-
Safety-gated Apple Notes MCP server and Codex plugin for macOS: local review, exact move plans, one-use authorization, journals, and rollback.
MIT -
Offensive-security tools, skills, and MCP servers for local AI agents — privacy-first, small-model-driven. Authorized pentest/red-team companion: 6 MCP servers (27 tools), all irreversible classes human-gated.
Apache-2.0 -
A clonable, self-contained starter for a governed Obsidian vault driven by an MCP server: domain ontology, rules registry, authority levels, and a writing-preflight gate.
MIT -
Authorization-gated penetration testing MCP server for AI agents - 80 tools (recon, vuln scanning, web app, exploit, CTF, reporting) with engagement-file scope enforcement. pipx install pentester-mcp
pypi MIT ★ 2 -
buzz-axirugithub.com/axiruai/buzz-axiruupdated this year
Bounded spend authority for Buzz agents. A gating MCP proxy in front of your payment MCP server: policy evaluation before money moves, human approvals in the Buzz channel, and a hash-chained decision log keyed to the agent's Nostr pubkey.
npm Apache-2.0 -
MCP server that gives local LLMs (LM Studio, Ollama) live web access plus recon, OSINT, and authorized security-testing tools like Burp/ZAP/Wapiti integration, APK triage, and self-extending skills, all safety-gated by default.
MIT ★ 3 -
retraction-watch-mcpgithub.com/handsomeZR-netizen/retraction-watch-mcpupdated this year
Local-first academic-integrity screening tool. Drag a PDF/DOCX/LaTeX, every reference checked against the Retraction Watch database with three-gate precision (title + year + author surname). Web app + CLI + MCP server, all sharing the same engine.
MIT -
A production-grade, self-hostable MCP server for Quera, bringing Quera’s problemset, submissions, courses, contests, LMS, AI, judge tooling, and authoring workflows to LLM agents. Built with async Python, 259+ tools, RBAC, safety gates, audit logs, revisions, multi-account support, and Docker deployment.
MIT -
Scope-gated, modular security MCP server — 205 tools across 26 skill packs (red/blue/OSINT/cloud/AD/post-ex/IoT/Wi-Fi/AI-LLM/mobile/forensics/CTF) with an attack-path engine, credential vault & ATT&CK reporting. For authorized engagements.
MIT -
One command installs 670+ security tools on Debian/Ubuntu/Kali, Fedora, Arch, openSUSE and Termux. Its authorization-gated MCP server runs them in a disposable Kata Containers VM sandbox for Claude Code, Codex, Gemini CLI, OpenCode and other MCP clients. 872 agent skills for CTF, pentesting, bug bounty, DFIR and red/blue teams.
oci MIT ★ 98 -
Agentic control layer for the Censurado self-hosted AI news portal: a stdlib Python CLI plus an MCP server (32 tools) an LLM agent drives to write and publish AI-persona articles through a gated editorial walk, curate the front page and authors, and deploy. Ships the prompt/skill recipe and a 24/7 serve loop.
MIT ★ 1 -
concordance-enginegithub.com/matharrismma/lighthouseupdated this year
Four-Gates packet validation engine with cross-disciplinary computational verifiers and MCP server for AI assistant integration.
pypi MIT -
Shotoku MCP server - gate any MCP tool call against a local spend and action policy.
npm MIT -
mcp-agentlockgithub.com/webpro255/mcp-agentlockupdated this year
An MCP server wrapper that gates tool calls on where a value came from, not what it says. Catches cross-hop prompt-injection laundering.
AGPL-3.0 -
wardergithub.com/santosh-singhh/warderupdated this year
A policy gate for MCP tool calls. Blocks destructive AI agent actions before they run — as the subprocess your AI client launches, where network gateways can't see.
cargo MIT -
Deterministic pre-execution safety gate for AI agents: verify tool calls against trusted evidence, policies, and WorldState before side effects.
★ 1 -
AI writes. SPARDA proves. A deterministic, offline gate that catches when an AI edit removes a guard, exposes a route, or breaks an invariant — no API key, right in the agent edit loop.
npm NOASSERTION ★ 9 -
Human approvals for agent actions. Gate refunds with Auth0-verified, exact-operation approval and resume them as MCP Tasks. Experimental TypeScript reference implementation with a hosted demo. Unofficial.
MIT -
AIC-gated command executor: project a command vector into a CLC operation, decide it against the caller's AIC grants, run it without a shell, and write signed evidence for every invocation. CLI + MCP entry points.
Apache-2.0