Development MCP Servers

Pre-trade token safety checks for Solana SPL and Robinhood Chain ERC-20 tokens: PASS/FLAG/FAIL verdict with authorities, LP lock, holders, liquidity, honeypot/tax. Pay per call ($0.01 USDC via x402 on Base or Solana); free quote/sample tools.

MCP server for coding agents building on Gemmein, the backend for web and mobile AI apps. It gives the agent the docs as tools and runs live isolation checks against the server-enforced access rules. Runs locally with npx -y @gemmein/mcp.

Reqode Coding MCP connects agents to requirements, architecture, mapped code and change context, with tools to create and resolve Findings.

MCP server for the Askell payment and subscription API. Discover endpoints, inspect customers, contracts, and billing, and call the API from Cursor or Claude.

Draw Hola Cloud gives AI agents a visual canvas to programmatically create architecture diagrams, system designs, workflows, sequence diagrams, flowcharts, infrastructure layouts, and other visual explanations, then return a shareable URL that can be opened, edited, presented, or shared.

Connect any AI agent to 1,000+ apps and 27,000+ actions through one remote MCP server.

Execution assurance, x402 compatibility checks and trust infrastructure for autonomous agents. Includes free diagnosis, paid Base/USDC compatibility receipts, agent registration and machine-readable discovery.

Signed agent identity, trust scoring, credit economy, and social layer for AI agents.

NarshaMCP is an Unreal Engine MCP server. It is distributed on Fab as NarshaADK. It parses C++ source, PDB symbols and binary .uasset files and serves the result over MCP. It is tested with Claude Code and Codex. Most reads work from disk with the Unreal Editor closed; edits need the Editor.

Local MCP server for Baton: edit git-native API requests, collections, environments and scripts.

Shared skills library for coding agents

Local-first multi-engine AI agent dashboard (Windows/macOS) with a built-in MCP server: expose agent sessions, long-term memory search, and 40+ tools to any MCP client. Runs Claude Code, Codex and a built-in ReAct engine side by side. https://github.com/phinn/KinetAios

Build production apps and manage Retool with AI agents. Retool MCP capabilities include react app building, workspace management, such as listing users, app permissions, resource and data exploration and user management.

Connects your AI assistant to 1,500+ certified utility APIs for document reading, verification, file conversion, and live data lookups.

Third-party offline stdio MCP for Basecoat UI composition in Astro/static HTML: curated templates, static validation, no runtime network.

Publish, test and share browser games on PlayDrop from your AI agent.

Independent security review for AI-built apps: exposed secrets, broken auth, unsafe data access.

SDK-free MCP server framework — self-implemented protocol engine (2026-07-28 + legacy), 19 plugins, OWASP security, zero dependencies

Autonomous coding pipeline: a frontier model plans and reviews, open-weight models write the code, gated by TDD and review.

An AAS MCP adapter that exposes configured Asset Administration Shell APIs as Model Context Protocol tools, enabling LLM agents to interact with any AAS-compliant backend.

Manage uCoz sites from an AI assistant: templates, content, shop, users, and files.

Search proxy setup guides, generate proxy configuration templates for curl, Requests, HTTPX and Playwright, and diagnose proxy errors such as a CONNECT 407.

Inspect monitoring coverage, propose configurations, and create uptime monitors, job heartbeats and webhook sources with scoped permissions.

Prior-art failure index for agents: pip install claimidx[server] → claimidx init --agent NAME → claimidx claim --yes (ask alone does not count). MCP: claimidx-mcp.

Local-first shared work-state for coding agents (Claude Code, Cursor, Codex, Gemini, Grok): presence, task claims, overlap warnings, live inter-agent messaging, ownership handoffs, and review packets.

A machine-enforced workflow protocol for AI coding agents that eliminates false progress. Tasks require verifiable evidence to close, while decisions, assumptions, and open questions persist in a versioned reasoning graph for seamless agent handoffs. Designed to keep deep, long-running projects coherent across sessions, agents, and months.

Prove end users to agents and apps: login-links, OIDC clients, and API keys over remote MCP

Lets Claude or any MCP client work inside Unreal Engine 5.3 to 5.8 and Blender 3.6 to 5.2, and reads the editor back after every change.

Shared app-testing runbook: read the plan, record results per platform, raise and fix issues. A fixed issue sends the check back for a retest rather than passing.

Local MCP stdio server for discovering coding agents, requesting work, exchanging messages, and tracking handoffs through a Scout broker; requires Bun 1.3+ and trusted local clients.